Skip to main content
Internet Archive's 25th Anniversary Logo

Full text of "Proceedings"

See other formats


\iOHQ^ 


?s.w 


M^^^O^ 


\0^^ 


AMERICAN    SOCIETY 


OF 


CIVIL  ENGINEERS 


January,    1908. 


PROCEEDINGS  =  VOL.  XXXIV— No.  1 


Pre: 


AMERICAN 
SOCIETY  OF 

CIVIL 
LENGINEERSj 

kFQUNDEOi 


:d  t 
^^         Z^-  U.  C/VIL 

ay 
VV,tuAM  P.  Morse 

Published  at  the  House  of  the  Society,  aao  West  Fltty-seventh  Street,  New  York, 
the  Fourth  Wednesday  of  each  Month,  except  June  and  July. 

CoDvriffhted  1908,  by  the  American  Society  of  Civil  Engineers. 

Entered  as  SeS?Class  Mat'te/at  the  New  York  City  Post  Office,  December  15th,  1896. 

Subscription,  $6  per  annum. 


Vol.    XXXIV.  JANUARY,  1908.  No.  1. 


AMEEICAN  SOCIETY  OF  CIVIL  ENGINEEES. 

INSTITUTED    185  3. 


PAPERS  AND   DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


CONTENTS. 

Papers :  page. 

Safe  Stresses  in  Steel  Columns. 

By  J.  R.  Worcester,  M.  Am.  See.  C.  E 3 

Effect  of  Earthquake  Shock  on  High  Buildings. 

By  R.  S.  Chew,  Assoc.  M.  Am.  Soc.  C.  E 8 

Discussions  : 

Invar  (Nickel-Steel)  Tapes  on  the  Measurement  of  Six  Primary  Base  Lines. 

By  Messrs.  J.  A.  Ockerson,  Horace  Andrews,  and  Noah  Cummings 16 

Municipal  Refuse  Disposal:  An  Investigation. 

By  Messrs.  W.  M.  Venable,  Albert  A.  Cary,  E.  H.  Foster,  B.  F.  Welton, 
C.  Herschel  Koyl,  Louis  L.  Tribus,  and  H.  Norman  Leask 21 

The  Reinforced  Concrete  Work  of  the  McGraw  Building. 

By  Messrs.  Guy  B.  Waite,  and  E.  P.  Goodrich ; 49 

Memoirs : 

Nathaniel  Henry  Hutton,  M.  Am.  Soc.  C.  E 63 


PLATES. 

I.    Measuring  Fort  Snelling  Base  Line  with  300-Ft.  Steel  Tape 17 

II.    Cinder  Concrete  Test  House 51 

HI.    The  Bon  wit-Teller  Building  and  the  Salvation  Army  Warehouse 53 

IV.     Reinforced  Concrete  Garage 55 


Vol.  XXXIV.  JANUARY,   1908.  No.    1. 


AMEEIOAN  SOCIETY  OF  CIVIL  ENQINEEES. 

INSTITUTED    1852. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  uot  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


SAFE  STRESSES  IN  STEEL  COLUMNS. 


By  J.  E.  Worcester,  M.  Am.  Soc.  C.  E. 
To  BE  Presented  February  19th,  1908. 


The  subject  of  a  proper  allowance  for  stresses  in  columns  has  bee^. 
treated  so  often  by  theorists  that  it  may  seem  as  though  no  more  could 
be  said  on  the  subject  without  danger  of  exhausting  the  patience  of 
the  engineering  profession,  but,  in  spite  of  all  the  theories,  the  prac- 
tice of  steel  designers,  as  shown  by  the  specifications  in  general  use 
to-day,  may  well  bear  further  consideration. 

The  reason  for  this  is  that  all  "rational"  column  formulas,  based 
on  the  elastic  properties  of  the  steel,  are  founded  on  considerations 
which  are  applicable  only  to  ratios  of  length  to  radius  of  gyration  far 
beyond  those  allowed  in  actual  construction.  It  is  known,  in  a  gen- 
eral way,  that  steel  in  compression  should  not  be  strained  as  high  as 
in  tension,  and  there  is  a  popular  impression  that  the  only  reason  for 
this  is  that  when  the  ratio  of  I  to  r  increases  above  0,  or,  at  most^ 
above  a  value  very  little  above  0,  the  strength  becomes  lessened  rapidly 
on  this  account;  but  there  has  been  a  growing  tendency  to  neglect  the 
fact  that,  even  in  very  short  columns,  there  is  not  the  same  unit 
strength  manifested  against  compressive  and  tensile  stresses. 

Note. — ^These  papers  arc  issued  before  the  date  set  for  presentation  and  discussion. 
Correspondence  is  invited  from  those  who  cannot  be  present  at  the  meeting,  and  may  be 
sent  by  mail  to  the  Secretary.  Discussion,  either  oral  or  written,  will  be  published 
in  a  subsequent  number  of  Procecdinqs,  and,  when  finally  closed,  the  papers,  with 
discussion  in  full,  will  be  published  in  Transactions. 


Pajiors.]  Qj^YY.   STRESSES   IN   STEEL   COLUMNS  3 

The  reason  for  this  difference  is  manifest  from  a  moment's  con- 
sideration. It  may  be  admitted  that,  within  the  elastic  limit,  the 
modulus  of  elasticity  is  practically  the  same,  whichever  way  the  metal 
is  strained — in  tension  or  compression.  One  may  even  go  further  and 
admit  that  the  elastic  limit  is  practically  the  same  for  both  stresses; 
but,  what  happens  after  the  elastic  limit  is  passed?  In  tension,  the 
member  merely  straightens  out — if  it  is  not  straight  to  start  with — 
and  stretches,  while  with  every  increase  in  length  comes  an  increase  in 
resisting  strength  until  the  ultimate  strength  is  reached,  the  final 
strength  being  nearly  twice  as  great  as  it  was  at  the  elastic  limit. 

In  compression,  however,  soon  after  the  elastic  limit  is  passed,  the 
column  will  cripple,  and  the  more  it  cripples  the  weaker  it  becomes. 
It  is  not  necessary  to  consider  the  ideal  conditions  of  exact  equilibrium 
in  the  resisting  power  of  a  section,  when  crippling  would  not  take 
place,  because  the  equilibrium — unstable  at  the  best — is  not  attainable 
in  practice.  On  the  other  hand,  it  must  be  admitted  that  the  ductility, 
which  is  of  such  great  advantage  in  tension,  is  not  present  to  an 
appreciable  amount  in  compression,  and  that  the  ratio  between  work- 
ing stresses  and  destructive  stresses  in  all  structures  depends  on  the 
compression  members,  and  not  the  tension,  when  anything  like  equal 
working  units  are  allowed  in  the  two. 

An  examination  of  the  results  of  tests  of  full-sized  columns  made 
by  Tetmajer,  Marshall,  Christie,  Bouscaren,  Strobel,  Lanza  and  the 
Watertown  Arsenal,  shows  strengths  of  wrought-iron  columns,  in  which 
the  I  -^  r  does  not  exceed  120,  of  from  16  000  to  43  000  lb.  per  sq.  in., 
and  for  mild  steel,  from  18  000  to  46  000  lb.  per  sq.  in.  By  far  the 
larger  part  of  these  range  between  22  000  and  34  000  lb.  for  iron,  and 
between  22  000  and  46  000  lb.  for  steel.  It  is  very  noticeable,  also, 
that  one  finds  results  of  more  than  28  000  lb.  with  the  longest  length 
and  less  than  this  amount  with  values  of  Z  -f-  r  as  small  as  30.  While 
the  axis  drawn  through  the  central  portion  of  the  group  of  these  ex- 
periments, when  plotted,  shows  some  inclination  toward  lower  values 
for  increased  length,  the  center  of  the  groiip  lies  at  about  30  000  lb. 
when  I  -^  r  =  90,  and  there  is  very  little  increase  in  strength  mani- 
fested in  the  tests  with  a  lesser  length  than  this.  It  is  apparent,  there- 
fore, that  if  the  compression  is  allowed  to  run  as  high  as  16  000,  the 
factor  between  working  stress  and  ultimate  will  not  exceed  2.  In 
tension  members,  on  the  other  hand,  the  corresponding  factor  is  nearly 


4  SAFE   STRESSES   IN   STEEL   COLUMNS  [Papers. 

4.  The  answer  to  this  argument  is,  of  course,  that  nobody  cares  what 
the  factor  between  working  strain  and  ultimate  may  be,  as  one  is 
really  interested  only  in  the  elastic  limit,  which  it  is  never  intended 
to  reach.  Is  it  not  time  to  call  a  halt  on  this  line  of  reasoning?  Have 
not  engineers  been  overconfident  in  their  ability  to  design  structures 
so  that  all  possible  contingencies  are  taken  into  account?  One  would 
not  willingly  make  use  of  a  material  in  tension  which  had  no  stretch 
beyond  the  elastic  limit,  yet  it  would  be  in  no  way  more  hazardous 
than  to  neglect  the  fact  that  such  is  the  case  with  compression  mem- 
bers, unless  a  greater  factor  below  the  elastic  limit  were  allowed  in 
these. 

The  history  of  the  development  of  the  column  formulas  used  in 
bridge  specifications  may  shed  a  little  light  on  the  way  in  which  unit 
strains  have  crept  up. 

The  adaptation  of  the  Gordon  formula  to  wrought-iron  columns  by 
Rankine  had  for  a  nimierator  36  000  lb.  That  is,  Rankine  recom- 
mended that  this  value  be  assumed  for  the  ultimate  strength  of 
wrought  iron  in  compression  of  short  columns.  The  earlier  specifi- 
cations for  railroad  bridges,  in  which  Rankine's  formula  was  used, 
recommended  7  500  or  8  000  in  the  numerator  when  10  000  was  used 
for  tension,  and  this  difference  between  the  numerator  of  the  com- 
pression formula  and  the  tensile  unit  has  been  retained  to  a  large 
extent  in  specifications  until  recently, 

Wlien  the  straight-line  formula  was  first  introduced,  it  was  recom- 
mended for  the  reason  that  a  straight  line  could  be  drawn  that  would 
coincide  very  well  with  the  plotted  results  of  experiments  for  ratios 
of  Z  to  r  between  90  and  150,  and  that,  in  giving  less  values  than  ex- 
periments warranted  above  this  point,  it  erred  on  the  side  of  safety. 
The  straight  line,  thus  drawn,  when  prolonged  the  other  way,  reached 
the  I  —-  r  =  0  line  at  about  the  tensile  value  of  the  steel,  making  the 

formula  take  the  form,  A  ^=  B  ■ —  C  X       in  which  A  =  the  allowable 

r' 

compressive  stress,  and  B  ^=^  the  allowable  tensile  stress.  This  simple 
form  appealed  strongly  to  engineers,  and  was  readily  accepted  by 
many,  but  the  fact  was  not  recognized  by  all  that  the  line  when  plotted 
goes  far  above  the  experiments  for  values  of  Z  -f-  /•  less  than  90.  The 
tables  by  C.  L.  Strobel,  M.  Am.  Soe.  C.  E.,  for  the  strength  of  Z-bar 
colunuis  were  bas(Ml  on  ;i  straight -line  fornnila,  but  this  is  a  notable 
instance  of  recogiiiiioii  of  tlic  error  of  the  formula   for  short  lengths, 


Papeis.]  SAYT.   STRESSES   IN*   STEEL   COLUMNS  5 

because  he  limited  his  stresses  to  12  000  when  the  straight  line  went 
higher  than  this  amount. 

At  this  point,  may  be  noted  what  seems  to  have  been  an  unwar- 
ranted change  in  specitications,  due  to  the  reprehensible  practice  of 
copying  from  one  to  another  with  slight  changes.  There  have  always 
been  many  engineers  who  liked  the  form  of  the  Eankine  formula  and 
refused  to  give  it  up.  Many  appear  to  have  been  struck  with 
the  simplicity  of  the  straight-line  formula  in  having  the  unreduced 
compression  unit  the  same  as  the  tension,  and,  wishing  to  take  ad- 
vantage of  this  feature,  but  still  adhering  to  the  Eankine  form,  they 
adopted  the  tension  unit  for  the  numerator  of  the  formula.  This 
throws  the  curve  entirely  above  the  field  of  tests,  and,  apparently,  can- 
not be  defended  by  any  reasoning. 

A  later  development,  of  the  specifications  which  are  based  on  the 
form  of  the  Rankine  formula  and  still  retain  the  tension  unit  in  the 
numerator,  is  to  adopt  a  lower  constant  in  the  denominator.  This, 
by  some,  is  made  20  000,  and  by  others,  8  000.  The  former  brings  the 
curve  within  the  outer  limits  of  the  group  of  tests,  while  the  latter 
passes  well  through  the  middle  of  the  group  for  values  of  Z  -=-  r  greater 
than  50,  but  is  above  the  group  for  lower  values. 

Perhaps  enough  has  been  said  to  show  that  the  formulas  in  general 
vise  to-day  need  to  be  sawed  off  at  the  end  toward  low  values  of  I  -f-  r. 
It  may  also  be  said  that  they  all  need  to  be  amputated  at  the  other  end. 
Mr.  Schneider,  years  ago,  suggested  that  values  of  Z  -^-  r  greater  than 
100  should  not  be  allowed  in  main  members,  and  this  limitation,  with 
slight  variations,  has  been  generally  accepted  since  that  time  as  an  es- 
sential of  good  practice. 

If,  then,  the  Rankine  formula  be  used,  with  the  numerator  value 
equal  to  the  tension,  and  the  compression  stress  be  limited  to,  say, 
75%  of  the  tension,  and  the  value  of  Z  -=-  r  to  100,  or  thereabouts,  one 
obtains  for  a  diagram  a  horizontal  line  running  to  a  cusp,  then  a 
concave  curve  running  to  another  cusp,  then  another  straight  line. 
Could  anything  be  more  irrational?  The  straight-line  formula  is  little 
better;  the  only  difference  being,  that,  in  the  middle  portion,  there  is 
a  straight  line  instead  of  the  curve.  How  much  better  it  would  be 
to  use  a  continuous  curve  throughout,  embodying  its  own  limitations 
at  each  end! 

The  late  J.  B.  Johnson,  M.  Am.   Soc.  C.  E.,  suggested  this  same 


G  SAFE   STRESSES  IN   STEEL   COLUMNS  [Papers. 

thought  in  his  book  on  Modern  Framed  Structures,  and  proijosed  a 
parabola.  This  is  safe  and  simple,  though,  if  the  vertex  is  kept  down 
to  a  safe  value  of  stress  for  short  lengths,  and  the  limitation  of  Z  -=-  r 
is  made  not  higher  than  120,  the  central  portion  of  the  curve  does 
not  reach  as  high  as  tests  would  warrant. 

An  elliptical  curve  fits  the  case  much  better,  the  ellipse  being  drawn 
with  its  center  at  the  zero  value  for  both  stress  and  I  -^-  ?%  and  having 
for  one  semi-diameter  the  limiting  value  oi  I  -^  r,  and  for  the  other 
tlio   limiting  stress  for  zero  lengths.     The  form  of  this  equation   is: 


B.n-    ' 


(Orf 

in  which  A  =  the  allowable  stress,  B  =  the  maximum  stress  at  I  ^-  r 
=  0,  and  C  =  the  maximum  value  of  Z  -^-  /■  allowed.  This  curve  is 
easy  enough  to  plot  as  an  ellipse,  but,  if  a  diagram  be  only  arranged 
so  that  on  the  scale  of  ordinates  B  is  of  the  same  length  as  C  on  the 
scale  of  abscissas,  the  curve  becomes  the  quadrant  of  a  circle. 

The  diagram,  Fig.  1,  illustrates  graphically  a  number  of  curves  of 
well-known  specifications,  together  with  the  results  of  tests,  by  the  ex- 
perimenters previously  referred  to,  reduced  so  as  to  allow  for  a  proper 
factor  of  safety.  This  reduction  is  made  so  that  the  experimental  re- 
sults can  be  compared  with  the  formulas  in  their  usual  forms.  Thi 
reduction  applied  is  proportional  to  the  ratio  between  the  tension  unit 
and  the  ultimate  tensile  strength  of  the  metal.     That  is,  for  wrought 

!•'>  000 
iron,  the  test  values  are  multiplied  bv  r^  n,  ,-,.  and  for  steel  the  multi- 

16  000 
i^^^^"  ^^   (50  000- 

The  proposed   fornuda,  as  plotted,  is  based  on  limiting  values  of 

compressive  stress  of  12  000,  and  oi  I  -^  r  of  120,  which  appear  to  be 
warranted  by  experiments  and  by  good  practice,  and,  as  the  scales  are 
arranged,  the  curve  is  circular. 

The  writer  puts  forward  a  new  formula  with  great  diifidence,  know- 
ing well  that  custom  is  a  very  difficult  thing  with  which  to  contend, 
and  how  cold  a  reception  new  compression  formulas  have  met  in  the 
liast;  but,  considering  how  poorly  the  formulas  now  in  use  fulfill  the 
requirements,  and  realizing  that  the  public  is  fully  awakened  at  the 
present  time  to  their  insufficiency,  the  time  seems  to  be  opportune  for 
at  least  suggesting  tlio  possil)ility  of  an  improvement. 


Papers.] 


SAFE   STRESSES   IN    STEEL   COLUMNS 


Vol.   XXXIV  JANUARY,  1908.  No.  1. 


AMEEIOAN  SOCIETY  OF  CIVIL  ENaiNEEES. 

INSTITUTED    1852. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  tlie  facts  and  opinions  advanced 
in  any  of  its  publications. 


EFFECT  OF  EARTHQUAKE  SHOCK  ON  HIGH 
BUILDINGS. 


By  K.  S.  Chew,  Assoc.  M.  Am.  Soc.  C.  E. 
To  BE  Presented  March  4tii,  1908. 


In  submitting'  this  paper,  the  writer  is  well  aware  that  he  is  deal- 
ing with  a  force  that  can  be  measured  only  by  the  resistance  en- 
countered; and  it  was  simply  with  a  view  of  determining  the  nature 
of  the  stresses  induced  in  structures  by  a  shock,  such  as  that  in  San 
Francisco  on  April  18th,  1906,  that  the  following  was  undertaken. 

All  realize  that,  with  a  possible  exception,  the  steel-framed  struc- 
tures in  San  Francisco  stood  this  shock.  This  fact  has  promoted  con- 
fidence, and  has  satisfied  architects  and  owners  that  such  is  the  safe 
type  of  building  for  the  Pacific  Coast.  The  engineer,  however,  cannot 
be  satisfied  until  he  ascertains  just  how  a  disturbance  of  this  nature 
affects  high  buildings. 

The  eft'ect  of  an  earthquake  is  to  produce  a  complex  movement  in 
the  crust  of  the  earth.  This  movement  is  a  wave  motion  accompanied 
by  more  or  less  twisting.  This  twisting  or  torsional  effect  is  small, 
and  affects  oidy  the  first  tier  of  columns.  The  length  of  the  wave  is 
very  long,  so  that  the  vertical  movement  is  small,  and,  for  a  stnu'ture 
with  a  well-designed  foundation,  may  be  neglected.  The  effect  of  the 
shock,  then,  is  from  the  horizontal  motion,  which  is  a  rapid  oscillation. 

Note. — These  papers  are  issued  before  the  date  set  for  presentation  and  discussion. 
Correspondence  is  invited  from  those  who  cannot  be  present  at  the  meeting,  and  may  be 
sent  by  mail  to  tlic  Secretary.  Discussion,  either  oral  or  written,  will  be  published 
in  a  subsequent  number  of  Proceed i)u)s.  and.  when  finally  closed,  the  papers,  with 
discussion  in  full,  will  be  published  in  Trans(i.etions. 


Papers.]    EFFECT  OF  EARTHQUAKE  SHOCK  ON  HIGH  BUILDINGS 


From  iiu'cluiiiit's,  it  is  known  that: 

Force  =  Mass  acceleration,  ov  F  =  M  a  = 


Wa 

9 


a) 


or,  the  force  of  the  earthquake  on  any  structure  is  the  mass  of  the 
structure  into  the  acceleration  produced. 


W 

,  1 1st  Floor 

9 
Fig.  1. 

Imagine  the  structure  represented  by  Fig.  1  to  be  built  of  a  per- 
fectly rigid  material,  and  that  W  is  the  weight  of  each  story;  then, 
as  the  foundation  takes  up  the  movement  of  the  earth,  it  endeavors  to 
set  the  structure  in  motion.     The  inertia  of  the  building  resists,  and 

5  Wa 


calls  into  play  the  force,  F 
floor  are : 


g 


S,=     4: 


>S3=8 


The  shearing  stresses  at  each 


Wa 

9 
Wa 

9 


S,=  2  ~  etc. 
9 

5  W  a  ^h 
The  maximum  bending  moment  =    X 


9 


2' 


If  these  shears  and 


bending  moments  could  be  developed,  the  building  would  follow  the 
movement  as  a  whole.  There  are,  however,  no  perfectly  rigid  materials, 
so  that,  under  the  action  of  a  force,  there  would  be  deformation  which, 
as  will  be  seen  later,  is  different  in  different  types  of  buildings. 

Consider  tirst  a  structure  which  has  no  wind  bracing.    By  reference 
to  Fig.  2  it  will  be  seen  that  each  story  weighs  W,  and  that,  there- 


K) 


EFFECT  OF  EARTHQUAKE  SHOCK  ON  HIGH  BUILDINGS   [Papers. 


fore,  the  resistance  that  each  story  would  offer  to  having  set  up  in  an 

acceleration,  a,  woiild  be . 

<J 

The   bvTilding,   under   this    action   of   forces,   is   a   beam   cantilever 
under  a  uniform  load  so  that : 

Moment  of  inertia  at  section  1  —  1  =  /^^ 

li  u  ii  a  li  9 o  ^  / 

assuming  I  =     '         J*   then,  approximately, 

-=T^'^,^z ^';> 

II  vpill  be  noticed  that     J,  in  Equation  2,  does  not  represent  the  dis- 
placement due  to  the  shock,  and  may  be  greater  or  less  than  this. 


F-  5  Wa 


t. 


g 


\  w 


w 


X 


Wa 

g 


Wa 

g 


Wa 

g 


Wa 

g 


Wa 

g 


Fig.  2. 

If  the  building  be  represented  as  a  single  line  it  can  be  seen  that,  as 
^  varies  directly  as  a,  Iv',  and  W ,  and  inversely  as  E  and  I,  that  either 
of  the  three  curves  may  be  attained,  dependent  on  these  variables. 

If  now,  E  and  7  be  large,  and  W  and  L  small,  and  the  oscillation 
rapid,  then  the  building  would  endeavor  to  follow  the  movement  closely, 
in  which  case  the  curves  produced  by  the  oscillation  would  produce  a 
wave  motion  in  the  structure  as  shown  approximately  by  Fig.  -i. 

The  writer  believes  this  to  be  the  effect  of  the  late  shock  as  felt  on 
the  majority  of  low  structures.     He  identifies  it  as  the  effect  produced 


Papers.]    EFFECT  OF  EARTHQUAKE  SHOCK  ON  HIGH  BUILDINGS 


11 


on  the  three-story  apartment  house  which  he  was  in  on  that  memorable 
morning.  It  was  quite  different  from  the  effect  felt  in  the  fifth  story 
of  a  six-story  steel-frame  structure  during  a  later  shock.  In  the 
latter  case,  there  was  a  decided  swinging  movement  of  the  building. 


Fig.  3. 
Referring  to  Fig.  1,  one  may  note  the  high  shearing  stresses  pro- 
duced.     These    are,    of    course,    the    same    in    Fig.    2.      The    bending 
moment  is  a  maximum  at  the  foot  of  the  column,  and  equals : 


5  Wa  h    ,    _  „^   J 
g       2  2 


2 


(t  +  ') 


(•■5) 


Original 
Eositii 


L 


■^^^^^^^^^^^^^^^^^^^^^^^^^^^55^? 


1st  Bloveuient        1st  Reversal 


2nd  Reversal 


3rd  Reversal 


Fig.  4." 
It  can  be  seen   at  a  glance  that  this  type  of  construction   is  not 
adapted  to  resist  any  shocks  except  those  of  very  small  displacement ; 
and,  even  in  these,  the  buildings  will  fail  in  §  ? 

detail.     For  instance:  It  was  noted,  after  the  fe|3 

shock  of  April  18th,  that  in  a  number  of  cases 
the  connection  of  beams  to  columns  had  failed 
by  the  rivets  shearing  off.  Reference  to  Fig. 
5  will  explain  the  condition.  By  referring  to 
the  curves,  it  will  be  seen  that  during  the 
vibration     the     column     bent,     throwing     the 


V\G. 


12  EFFECT  OF  EARTHQUAKE  SHOCK  ON  HIGH  BUILDINGS  [Papers. 

couple,  F  F,  into  action,  and  the  rivets,  not  being  sufficient  to  stand 
this,  failed.  The  writer  has  also  noted  several  buildings  of  this  type 
in  which  there  is  a  decided  crack  in  the  brickwork  following  the 
column,  which  tends  to  substantiate  this  theory. 

A  wind-braced  building  will  act  a  little  differently  from  the  fore- 
going, due  to  the  fact  that  the  point  of  contraflexure  in  the  columns 
is  fixed  by  the  bracing,  so  that  the  building  in  part  will  follow  the 
movement. 

Let  D  =  the  horizontal  displacement, 

t  =  the  time  for  said  horizontal  displacement, 

a  =  the  acceleration  ^  — „^ 

g  =  the  effect  of  gravity; 

W  a  WD 

then,  the  force  exerted  on  the  building  ^  F  ^   —    ,    where 

9  0  t- 

W  equals  the  weight  of  the  building. 

If,  in  Fig.  6,  it  be  assumed  that  the  horizontal  girders  are  stiff 
enough  to  fix  the  columns  at  the  knees,  then  the  effect  on  the  build- 
ing by  the  movement,  D,  is  as  shown. 

D'     =  Aj   +  A2 -\-  h  =    2    (^1  +    ^2 +  ^5)  =  the  de- 
formation in  building. 
Wr,  =  the   weight   of   the   building   above  and   including  the  first 

floor. 
W^  =  the  weight  of  the  building  above  and  including  the  second 

floor. 
W^  =  the  weight  of  the  building  above  and  including  the  third 

floor. 
From  mechanics,  it  is  known  that: 
_      Fl^    _  WD^JP_ 

~    :iE  I  ~    gf   -.IE  I ^   ^ 

W.  1?.  T) 


•i  E  I 

w,  i\ 

B 

'.iEI^ 

yt^ 

W,  l\ 

D 

^s     =         ^  ^  r„   etc 

*  'd  E  I^f]  t- 


Papers.]   EFFECT  OF  EARTHQUAKE  SHOCK  ON   IIIUII  BUILDINGS 


13 


The  bending  moment  is  a  maximum  at  the  base,  and 
_    W^D  I,  W,'  I,  D 


gt' 


\    ^  2,  El) 


(") 


Fig.  6. 


These  conditions  are  reached  for  the  movement  in  one  direction. 
As  this  movement  is  back  and  forth,  it  gives  the  approximate  curves 
shown  by  Fig.  7. 

It  can  readily  be  seen  that  while  the  curve,  0  A,  shows  the  curve 
in  the  building  for  the  movement  1  —  A,  that,  before  the  return  move- 
ment throws  the  reverse  curve  0  —  B  into  the  structure,  a  portion 


14  F.Fl'KCT  or  EAItTHQUAKE  SHOCK  ON  HIGH  BUILDINGS   [Papers. 


of  the  frame  at  the  top  will  endeavor  to  straighten,  or  that  the  point, 
0,  will  move  to  .r,  and  that  the  curve  on  the  beginning  of  the  return 
movement  will  ha  x  y  z  A.  '  This  tendency  is  aggravated  on  each  re- 
verse, and  produces  the  whip  action  at  the  top. 


Fig.  8. 


Fig.  7. 

By  Equation  5  it  may  be  seen  that  should  D  be  very  small  and  L 
on  the  first  story  large,  nearly  all  the  bending  would  occur  in  the  first- 
story  columns,  the  building  above  receiving  a 
very  small  force.  In  this  case,  the  first-story 
columns  vibrate  back  and  forth,  and  the  build- 
ing above  is  practically  stationary.  Of  course, 
this  would  be  productive  of  a  high  bending  move- 
ment in  the  first-story  columns,  and  a  shock  of 
any  magnitude  would  wreck  the  building. 

If  the  foregoing  analysis  is  correct,  the  following  may  be  noted: 

1st. — That  the  stresses  produced  are  similar  to  those  caused  by 
wind ; 

2d. — That,  on  account  of  quick  reversal,  the  stresses  are  increased; 

3d. — That,  in  a  wind-braced  structure,  the  total  effect  is  distributed 
throughout  the  structure; 

4th. — That,  as  this  effect  is  a  direct  function  of  the  weight,  the 
wall  and  floor  construction  should  be  as  light  as  consistent  with 
strength ; 

5th. — That,  as  this  effect  is  inversely  as  the  coefficient  of  elasticity, 
the  frame  should  be  of  a  highly  elastic  material; 

Gth. — That,  as  the  effect  in  buildings  that  are  not  wind-braced 
varies  as  the  cube  of  the  height,  these  structures  should  be  limited 
in  height; 

'i'th. — Tliat  a  inouolitliio  fonudation  is  preferable  to  one  having 
isohited  footings. 


Papers.]    kffecT  OF  EAltTIKJUAKE  SHOCK  ON   IIKill  BUJLDINGS  15 

These  conclusions  all  point  to  a  steel  frame  with  reinforced  walls 
and  floors  as  the  type  of  construction  for  the  vicinity  of  San  Francisco. 

With  respect  to  reinforced  concrete,  the  writer,  although  he  be- 
lieves it  to  be  a  valuable  combination,  thinks  it  unsuited  to  resist 
the  forces  that  an  earthquake  shock  would  produce  in  a  high  building, 
for  the  following  reasons : 

1st. — This  type  of  construction  is  not  adapted  to  resisting  reversed 
stresses ; 

2d.- — It  cannot  take  shock  ; 

3d. — The  construction  is  heavy,  which  conflicts  with  Conclusion  4; 

4th. — The  coefficient  of  elasticity  is  low,  which  does  not  agree  with 
Conclusion  5; 

5th. — The  high  bending  moments  produced  in  columns  and  girders 
would  make  their  designs  uneconomical; 

6th. — Added  to  these,  when  it  is  considered  that  any  failure  in 
detail  will  necessitate  the  renewal  of  several  entire  members,  the  dis- 
advantages of  this  construction  will  be  seen. 

Finally,  the  writer  would  recommend: 

I.^ — The  building  to  have  lattice  girders  of  the  Warren  type,  as 
deep  as  the  spandrel  section  will  allow,  running  entirely  around  the 
structure  at  every  floor.  The  advantage  of  this  construction  is 
obvious :  Being  in  the  center  of  the  wall,  the  brickwork  or  concrete 
can  be  built  around  the  members,  the  wall  thereby  being  reinforced, 
the  girder  can  be  designed  economically  for  the  different  floors.  Be- 
ing deep,  it  forms  the  lintel  over  the  windows  and  at  the  same  time 
decreases  the  length  of  the  column. 

11. — Monolithic  foundations. 

III. — Reinforced   concrete    walls,    the   reinforcement   to    run    hori 
zontally  and  vertically. 

IV. — Floor  slab  to  be  of  rock  concrete  at  least  3J  in.  thick. 

V. — Floors  cut  up  by  a  large  number  of  openings  to  be  braced  so 
that  they  can  transmit  all  horizontal  shear  to  the  columns. 

VI. — Wind  bracing  connections  to  be  designed  to  develop  the  main 
member. 

VII. — Wind  bracing  to  be  carried  to  the  ground. 

VIII. — Columns  to  be  calculated  for  an  extreme  fiber  stress  of 
12  000  lb. 


Vol    XXXIV.  JANUARY,  1908.  No.  1. 


AMEEIOAN  SOCIETY  OF  CIVIL  ENGINEEES. 

INSTITUTED     18  5  2. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


INVAR  (NICKEL-STEEL)  TAPES  ON  THE 

MEASUEEMENT  OF  SIX  PRIMARY 

BASE  LINES. 

Diyciissiou.* 


By  Messrs.  J.  A.  Ockerson,  Horace  Andrews,  and  Noah  Cummings. 


Mr.  Ockerson.  J.  A.  OcKERSON,  M.  Am.  Soc.  C.  E.  (by  letter). — Mr.  French's  ac- 
count of  the  progress  made  in  the  use  of  steel  tapes  in  geodetic  work 
is  interesting  and  valuable.  The  old-time  methods  of  using  bars  or 
rods  for  the  measurement  of  base  lines  were  both  laborious  and  ex- 
pensive, and,  as  a  consequence,  the  intervals  between  such  lines  in  a 
system  of  triangulation  were  entirely  too  long. 

Professor  Woodward's  paper  on  long  steel  tapes,  and  the  discus- 
sion thereon,t  gave  an  account  of  the  use  of  steel  tapes  in  connection 
with  the  triangulation  work  of  the  Coast  Survey. 

The  official  reports  of  the  Mississippi  and  Missouri  River  Commis- 
sions, of  earlier  date  by  several  years,  gave  accounts  of  the  steel-tape 
work  on  their  respective  surveys,  which  included  high-grade  triangu- 
lation, where  the  length  of  the  triangle  sides  and  the  closure  required 
were  such  as  to  compare  favorably  with  so-called  primary  work. 

In  August,  1880,  the  writer  made  use  of  a  steel  tape  in  the 
measurement  of  a  base  line  6  063  ft.  long,  opposite  Grafton,  111.,  in 
connection  with  the  triangulation  in  that  vicinity,  and,  although  the 
(M|uipment  was  deticient  in   many  respects,  the  results  obtained  were 


♦This  discussinn  (of  the  paper  by  Owen  B.  French,  M.  Am.  Soc.  C.  E.,  printed  in 
Proceedings  for  October,  1907),  is  printed  in  Proceeding'^  in  order  that  the  views  expressed 
may  be  brought  before  all  members  for  further  discussion. 

f  Transactions,  Am.  Soc.  C.  E.,  Vol.  XXX,  pp.  8M0r  and  638-652. 


PLATE  I. 

PAPERS,   AM.  SOC.  C.   E. 

JANUARY,   1908. 

OCKERSON  ON 

STEEL  TAPES. 


1  *ii*-^1, 


Fig.  1.— Measuring  Fort  Snelling  Base  Line  with  3G0-Ft.  Steel  Tape, 
Rear  End  of  Tape. 


Fig.  2.— Front  End  of  Tape,  Fort  Snelling  Base  Line,  Showing  Tension  Device. 


Papers.]  DISCUSSION   ON   INVAR    (nICKEL-STEEL)    TAPES  17 

STich   as  to  establish  practically  the  use  of  the  steel  tape  in  the  ex-  Mr.  Ockerson. 
tensive   triangulation   work   which   followed,    from   the   mouth   of   the 
Ohio  to  the  headwaters  of  the  Mississippi. 

The  writer  believes  this  to  be  the  first  use  of  the  steel  tape  in  re- 
fined geodetic  work,  at  least  in  the  United  States. 

The  methods  of  manipulation  in  the  field  were  modified  and  im- 
proved, in  the  interest  of  both  economy  and  accuracy,  as  experience 
developed  the  defects. 

Mr.  O.  B.  Wheeler,  of  the  Missoviri  River  Commission,  is  entitled 
to  the  credit  of  introducing  an  accurate  tension  adjuster,  which  is 
described  in  the  Annual  Report  of  that  Commission  for  1886. 

The  use  of  metal  strips  on  which  to  mark  the  graduated  extremi- 
ties of  the  successive  tape  lengths  was  also  developed  in  the  river  sur- 
veys. This  was  a  very  important  step,  as  it  virtually  permitted  the 
graphical  results  of  each  tape  length  to  be  transferred  from  the  field 
to  the  ofiice,  where  the  discussion  could  be  taken  up  at  leisure. 

Mr.  Marshall,  in  connection  with  the  survey  of  the  Red  River, 
made  a  number  of  improvements,  among  which  was  the  use  of  two 
tapes  of  different  metals  at  one  and  the  same  time. 

The  greatest  source  of  error  in  the  use  of  the  steel  tape  lies  in 
failure  to  secure  its  temperature,  as  it  is  much  more  sensitive  to 
changes  than  the  thermometer  used  in  connection  therewith.  The 
writer  had  in  mind  a  method  of  diminishing  the  difference  between 
the  two,  by  the  construction  of  a  thermometer  with  an  elongated 
bulb  of  the  same  material  as  the  tape,  and  perhaps  extending  the 
metal  along  the  back  of  the  glass  scale  tube.  The  invar  tape,  ap- 
parently, eliminates  much  of  the  objection  to  the  steel  tape  incident 
to  changes  of  temperature. 

In  the  Mississippi  River  triangulation,  the  measurement  of  base 
lines  became  so  easy  that  the  general  practice  called  for  a  base  line  at 
intervals  of  about  12  triangles.  That  is  to  say,  the  instrumental 
errors  of  centering  both  instrument  and  target,  and  errors  of  point- 
ing, were  larger  than  the  errors  of  base  measurement,  hence  such 
errors  were  largely  localized  by  the  use  of  frequent  bases. 

In  the  writer's  opinion,  each  tape  should  be  standardized  by 
measuring  a  primary  base  the  length  of  which  has  been  determined 
by  a  Repsold  or  other  refined  base-measuring  apparatus.  The  measur- 
ing should  be  done  under  conditions  and  by  methods  identical  with 
those  to  be  used  in  the  measurement  of  a  new  base,  in  preference  to 
relying  on  a  laboratory  determination  of  the  length  of  the  tape. 

Table  9  gives  some  results  of  base-line  measurements  with  steel 
tapes  on  the  Mississippi  River  Triangulation.  The  measurements  were 
generally  made  in  the  morning,  before  sunrise,  when  changes  of 
temperature  were  not  very  rapid.  The  lengths  are  given  in  round 
numbers,  omitting  the  decimals. 


18 


DISCUSSIOX    ox    INVAU    (KICKEL-STEEL)    TAPES  [Papers. 


Mr.  Ockerson    TABLE    9. —  SoME    RESULTS    OF    BaSE-LiNE    MEASUREMENTS    WITH    StEEL 

Tapes  on  the  Mississippi  River  Triangulation. 


Mr.  Andrews. 


Location. 

Length  of  base  line,  in  feet. 

Discrepancy  between  suc- 
cessive measurements. 

New  Boston 

18  066 
5  624 
7  105 
7  00] 

5  312 

6  486 

5  223 

6  783 
5  379 
5  400 
5  401 
5  700 
5  400 
4  798 

1  in      759  000 

Rapid  City 

594  917 

East  Dubuque 

346  965 

Cassville 

266  412 

Prairie  du  ( 'hien 

"          265  000 

DeSoto 

929  300 

Trempeleau 

'•      2  396  000 

Wabasha 

740  000 

Red  Wing 

"      8  400  000 

Fort  Snellin^ 

517  000 

Monticello 

"      1  475  000 

Rice  

"      2  969  000 

Brainard 

Aitkin 

438  400 
'•      1  304  000 

The  method  of  handling  the  tape  is  shown  by  the  photographs  on 
Plate  I.  Single  measurements  of  lines  1  mile  long  have  been  made 
in  28  min.  In  the  lines  cited,  no  eifort  was  made  to  secure  a  very 
high  degree  of  accuracy,  but  simply  to  keep  within  the  limit  of  dis- 
crepancy between  two  measurements,  1  in  250  000,  as  prescribed. 

Horace  Andrews.  M.  Am.  Soc.  C.  E.(by  letter). — The  engineer- 
ing profession  is  indebted  to  Mr.  French  for  his  clear  and  useful 
exposition  of  the  practical  adaptability  of  invar  to  field  use. 

There  would  seem  to  be  little  left  to  be  desired  in  base-measuring 
apparatus,  now  that  temperature  corrections  are  so  well  eliminated. 
The  history  of  base  measurement  has' been  one  of  constant  struggle 
against  the  uncertainty  of  temperature  corrections.  At  present,  the 
use  of  iced  bars,  steel  tapes,  and  night  work,  enables  high  precision 
to  be  attained,  together  with  speed  and  economy  passing  all  former 
experiences.  A  further  and  most  important  advance,  from  the 
economical  standpoint,  is  now  assured  through  the  use  of  this  wonder- 
ful alloy. 

Previous  to  the  six  base  lines  referred  to  by  Mr.  French,  some 
35  base  lines  of  primary  importance  had  been  measured  in  the  United 
States.     The  three   earliest,  one   in   1834  and  two   in   1844,  measured 

1 


with  the  Hassler  apparatus,  had  an  average  probable  error  of 

Between  1847  and  1873,  seven  bases  were  measured  by  the  Coast  Sur- 
vey,   with    the    Bache-Wiirdemann    apparatus,    an    average    probable 

error    of     -__-_--—    being   indicated.       Similai 


4:57  000 

States 


United 
l>r()bal)l('   error  o 


..ake    Survey, 
f   five   base; 


from 
1 


ajiparatus    used     l)y    the 
1870    to    1875,    gave    the    average 
Till'   lu'psold   apparatus  then 


rapois.]  DISCUSSION    ON    INVAR    (nICKEL-STEEL)    TxVPES  19 

canio  into  use  on  the  Lake  Survey,  three  bases  being  measured  with  Mr.  Andrews, 
an  average  probable  error  of  oi^^y  wv7c~/)nA'     The  United  States  Coast 

ajid  Geodetic  Survey,  after  1873,  measured  eight  bases  with  various 
apparatus.  Two  of  these,  measured  in  1891-92  with  the  iced  bar  and 
steel   tape,    showed   excellent   residts.     The   average  probable   error   of 

these  eight   bases  was  ^,^-  „„^.     Then  came  the  phenomenal  achieve- 

nient  of  1900,  when  the  United  States  Coast  and  Geodetic  Survey, 
having  commissioned  one  field  party  to  measure  nine  bases  in  a  work- 
ing season,  the  aim  being  to  secure  a  precision  of  about  rnn  ofin' 
obtained,   not   only   unprecedented   economy  of  time   and   money,   but 

an   average  probable  error  of   only  -i  i  en  aaa-     This   success   was   due 

to  the  use  of  the  iced  bar  and  steel  tapes,  as  mentioned  by  Mr.  French; 
the  advantages  of  invar  are  those  pointed  out  by  him,  and  are  irre- 
spective of  the  higher  precision  which  was  incidentally  obtained. 

Obviously,  invar  will  be  an  admirable  material  for  precision  level- 
ing-rods.  Its  use  for  pendulum  rods  was  one  of  the  first  suggested. 
A  pendulum,  supported  by  an  iron  rod,  will  change  its  rate  about  1 
mill,  a  week,  if  subjected  to  a  change  of  temperature  of  30°  fahr,, 
but,  with  invar  having  the  coefficient  given  by  the  author,  the  change 
of  rate  would  be  only  2  sec. 

It  would  be  interesting  to  know  the  exact  proportions  of  nickel 
and  steel  entering  into  the  composition  of  the  invar  tapes.  In  view 
of  the  fact  that  the  coefficient  of  expansion  given  in  Table  2,  is  only 
one-half  that  found  in  Guillaume's  36%  of  nickel  alloy,  it  would  seem 
that  some  change  must  have  been  made  in  the  proportions,  and  if  the 
name,  "invar,"  is  to  be  adopted,  it  would  be  well  to  have  it  apply  to  a 
definite  nickel-steel  alloy;  at  present,  there  are  two  "invar"  alloys, 
one  of  which  has  only  half  the  invariability  of  the  other. 

Engineering  measurements  in  general  must  be  made  under  all  con- 
ditions of  temperature,  and  it  will  be  of  great  advantage  to  be  inde- 
pendent of  temperature  corrections.  The  writer  has  found  it  very 
advisable  to  keep  temperature  notes  for  important  steel-tape  measure- 
ments, and  correct,  under  the  rule  of  0.01  ft.  in  100  ft.,  for  each 
15°  fahr.  In  the  surveys  for  the  Boston  Back  Bay  tract,  it  has  been 
stated  that  brass  tapes  were  used,  and  with  the  correction  of  0.01  ft. 
in  100  ft.  for  10°  fahr.  change  of  temperature.  With  invar,  the  cor- 
rection would  appear"  to  be  0.01  ft.  in  100  ft.  for  each  44°  fahr.,  so 
that  temperature  corrections  would  in  general  be  negligible. 

Noah  Cummings,  Assoc.  M.  Am.  Soc.  C.  E. — The  low  coefficient  Mr.cummings. 
of  expansion  of  invar  makes  it  a  most  desirable  material  for  a  tape, 


20  DISCUSSION   ON  INVAR    ( NICKEL-STEEL )    TAPES  [Papers. 

.Mr.Cummings.  and  the  results  given  in  the  paper  show  that  it  has  proved  very  satis- 
factory for  base-line  measurements. 

It  could  be  used  to  advantage  for  measuring  base  lines  for  city  or 
bridge  triangulations,  or  wherever  great  accuracy  is  required  in  base- 
line work.  However,  there  seem  to  be  serious  objections  to  the  use 
of  the  invar  tape  for  ordinary  city  surveying,  even  though  its  low  co- 
efficient of  expansion  would  practically  eliminate  temperature  correc- 
tions. It  is  more  easily  bent  and  less  elastic  than  steel,  and,  accord- 
ing to  the  makers,  requires  a  reel  16  in.  in  diameter.  Thus  great 
care  must  be  taken  in  handling  it,  and  it  would  need  to  be  retested 
every  time  someone  happened  to  run  into  it  while  a  measurement  was 
being  made  or  in  case  it  was  stepped  on,  either  of  which  may  easily 
happen  in  ordinary  street  surveying. 

It  is  a  question  whether  surveying  needs  to  be  done  more  accurately 
than  the  holding  of  the  points  established  by  the  survey.  When  street 
corners  are  marked  by  stone  monuments  placed  near  the  surface  of 
the  sidewalk,  there  is  a  possible  movement  due  to  frosts  and  to  ex- 
cavations for  building  and  street  construction  purposes.  If  the  invar 
tape  should  be  developed  so  as  to  be  more  easily  handled  and  then 
come  into  use  for  city  surveying,  a  greater  degree  of  accuracy  would 
be  obtained;  but,  to  secure  the  benefits  of  this,  the  points  established 
would  have  to  be  marked  on  stone  or  concrete  monuments  either  be- 
low or  extending  below  the  frost  line. 


Vol.  XXXIV.  JANUARY,  1908.  No.  1. 


AMERICAN  SOCIETY  OP  CIVIL  ENGINEERS. 

INSTITUTED    1853. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  nr.t  responsible,  as  a  body,  for  the  tacts  and  opinions  a,dvanced  in 
any  of  its  publications. 


Ml^NTCIPAL  REFUSE  DISPOSAL: 

AN  INVESTIGATION. 

Discussion.* 


By  Messrs.  W.  M.  Venable,  Albert  A.  Gary,  E.  II.  Foster,  B.  F. 

Welton,  C.  IIerschel  Koyl,  Louis  L.  Tribus,  and 

H.  Norman  Leask. 


W.  M.  Venable,  M.  Am.  Soc.  C.  E.  (by  letter). — In  this  valuable  Mr.  Venabie. 
paper  the  author  proves  that  the  refuse  of  the  Borough  of  Richmond, 
New  York  City,  contains  sufficient  calorific  material  to  enable  it  to  be 
burned  without  offense,  and  without  using  auxiliary  fuel.  He  also 
presents  data  regarding  forty  incinerating  plants  in  Great  Britain, 
with  the  object  of  determining  the  best  features  of  design  for  use  in 
a  proposed  plant.  The  investigation  which  led  to  the  conclusion  that 
it  is  desirable  to  burn  all  the  refuse  in  one  incinerator,  if  that  is  found 
practicable,  is  not  given  in  the  paper,  nor  is  there  any  investigation  of 
the  merits  or  demerits  of  incinerators  of  American  design.  The  writer 
is  of  the  opinion  that,  in  the  United  States,  it  is  seldom  desirable  to 
reburn  all  refuse,  including  the  ashes  from  private  houses  and  other 
buildings,  and  would  like  the  author  to  present  in  detail  the  data  upon 
which  this  determination,  which  preceded  the  investigation  reported 
in  the  paper,  was  based. 

Whether  or  not  the  method  of  destroying  all  wastes  in  one  set  of 
furnaces  will  be  found  the  best  for  municipalities  generally,  engineers 
are  indebted  to  Mr.  Fetherston  for  his  thorough  work  in  ascertaining 
the  quantities  of  garbage,  ashes  and  rubbish,  and  their  calorific  value, 
in  what  may  be  taken  as  a  representative  district.     It  is  remarkable, 

*  This  discussion  (of  the  paper  by  J.  T.  Fetherston,  Assoc.  M.  Am.  Soc.  C.  E.,  printed  in 
Proceedings  for  November,  1907),  is  piinted  in  Proceedings  in  order  that  the  views 
expressed  may  be  brought  before  all  members  for  further  discussion. 


23  DISCUSSION  ON  MUNICIPAL  liEFUSE  DISPOSAL  [Papers. 

Mr.  Venable.  also,  that  the  best  summary  of  British  practice  in  refuse  disposal  is 
found  in  this  paper,  by  an  American  engineer,  for  nse  in  America. 
Too  much  praise  can  hardly  be  given  for  the  judgment  shown  in  the 
preparation  of  the  various  tables,  although  Table  1  was  prepared  so 
long  ago  as  to  make  it  necessary  for  the  reader  to  guard  himself 
against  the  error  of  assuming  that  it  contains  all  the  data  now  avail- 
able on  the  subject  with  which  it  deals.  The  work  of  Messrs.  H.  de 
B.  Parsons,  Rudolph  Hering,  W.  F.  Morse,  and  others  has  been  pub- 
lished since  1904. 

From  this  paper  and  other  available  data,  it  is  safe  to  assume  that, 
in  almost  any  municipality,  if  all  the  household  refuse  is  collected  and 
brought  to  one  place,  the  mixture  will  contain  sufficient  calorific 
energy  to  make  it  practicable  to  burn  it  without  admixture  of  other 
fuel,  and  to  permit  the  generation  of  some  steam  for  power  purposes 
from  the  heat  in  the  gases  of  combustion.  It  does  not  follow  from 
this,  however,  that  such  collection  and  disposal  is  the  most  advisable. 
It  can  hardly  be  granted,  as  a  general  proposition  in  cities,  that  it  is 
impossible  to  collect  ashes  in  such  condition  that  sanitary  disposal  of 
them  without  reburning  is  impracticable.  If  such  is  granted,  as  ap- 
pears to  have  been  done  in  the  present  case,  there  would  still  be  rea- 
sons for  considering  separate  collection  and  burning  in  separate  parts 
of  an  incinerator,  keeping  ashes  separate  from  garbage  and  refuse, 
both  for  sanitary  reasons  and  for  convenience  and  economy  in  actual 
burning. 

If  the  reburning  of  ashes  is  to  be  decided  from  considerations  of 
economy  only,  it  should  be  regarded  entirely  apart  from  the  disposal 
of  other  wastes,  for  the  introduction  of  ashes  into  the  garbage  makes 
their  disposal  much  more  costly  than  otherwise,  even  if  it  is  neces- 
sary to  furnish  a  considerable  quantity  of  coal  to  assist  in  destroy- 
ing the  garbage. 

While  ashes  from  household  fires  contain  much  combustible  ma- 
terial, they  do  not  contain  enough,  as  a  rule,  to  make  up  for  the  cost 
of  stoking  them  through  a  crematory,  not  including  plant  charges; 
and,  unless  a  very  great  reduction  in  weight  is  secured  by  reburning, 
there  will  be  no  saving  in  total  haul  by  the  burning  process.  Gen- 
erally, the  weight  of  ashes  passed  through  a  crematory  is  not  very 
greatly  reduced,  although  the  weight  of  rubbish  and  garbage  is  very 
much  decreased  by  burning.  There  may  be  cases,  however,  where  a 
furnace  can  be  located  at  the  center  of  a  district,  and  the  haul  to  the 
dump  is  very  much  longer  than  that  to  the  furnace,  in  which  cases  the 
saving  in  haul  will  more  than  counterbalance  the  cost  of  dumping, 
stoking,  interest  and  dei)reciation  on  plant,  and  reloading  for  haul  to 
the  dump.* 

*  This  matter  is  discussed  in  the  writer's  book,  "Garbage  Crematories  in  America," 
John  Wiley  &  Sons,  1906.  In  this  book  will  also  be  found  descriptions  of  every  type  of 
crematory  installed  in  the  United  States,  reference  to  every  United  States  patent  o"f  interest 
in  this  field,  and  a  list  of  the  more  important  and  representative  plants  installed  by  each 
builder  of  such  works  in  tiie  United  States. 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  23 

In  the  United  States  it  has  been  customary  to  dispose  of  ashes  Mr.  Venabie. 
separately  from  garbage,  from  motives  of  economy,  and  furnaces  for 
the  disposal  of  garbage  or  refuse,  or  of  both  combined,  have  been  de- 
signed with  the  expectation  that  ashes  would  be  excluded.  It  is  prac- 
ticable to  burn  these  materials  properly  without  forced  draft,  and 
several  builders  of  crematories  have  accomplished  this  successfully, 
at  prices  of  disposal  per  ton  quite  as  low  as  those  obtaining  in  England 
for  the  mixed  refuse;  but  crematories  operating  on  natural  draft  can 
be  abused  more  readily  than  those  using  forced  draft,  and,  conse- 
quently, when  handled  by  the  ignorant  persons  who  are  so  often  placed 
in  charge,  the  furnaces  have  received  the  blame  that  ought  to  have 
been  charged  against  the  persons  in  authority.  Of  course,  very  many 
crematories  of  poor  design,  and  crematories  attempting  to  burn  ma- 
terials for  which  they  have  not  been  fitted,  have  been  installed,  and  ' 
the  blame  has  not  always  been  with  the  operator.  Crematories  of  the 
so-called  American  design  are  much  cheaper  to  build  than  those  of  the 
British  type,  as  they  require  no  boiler  plant,  or  power  auxiliaries. 
They  will  consume  successfully  garbage  and  rubbish  of  a  character 
which  cannot  be  burned  in  those  of  the  British  type,  and  are  very 
economical  in  the  use  of  labor  in  stoking.  Therefore  they  ought  not 
to  be  condemned,  or  left  out  of  consideration  in  selecting  a  method  of 
disposal,  but  should  be  installed  where  economy  shows  that  they  will 
be  most  economical  in  the  long  run ;  and  proper  precautions  should  be 
taken  that  they  are  operated  so  as  not  to  produce  a  nuisance. 

On  the  other  hand,  it  is  practically  impossible  to  burn  ash-bin 
refuse  with  natural  draft.  The  reason  is,  not  that  a  strong  enough 
natural  draft  cannot  be  obtained,  but  that  the  constant  opening  of 
doors  for  stoking,  on  account  of  the  large  proportion  of  ashes  to  actual 
fuel  in  the  mixed  refuse,  admits  to  the  furnace  too  much  air  for 
proper  combustion.  This  reduces  the  draft  and  also  causes  the  pro- 
duction of  foul  odors  in  the  chimney  gases.  It  requires  so  much  more 
head  to  create  a  proper  draft  through  a  mixture  of  ashes  than  through 
a  mixture  of  rubbish  that  it  is  possible  to  burn  the  rubbish  without 
offense,  on  natural  draft,  even  with  doors  frequently  open,  although  it 
is  not  possible  if  a  large  proportion  of  ashes  is  introduced  into  the 
mixture  to  be  burned.  Thus,  practically  all  British  incinerator 
builders  have  been  compelled  to  adopt  forced  draft  because  they  reburn 
ashes,  and  to  install  boilers  in  order  to  develop  power  to  obtain  it. 
When  forced  draft  is  used,  the  stack  should  be  designed  merely  to  carry 
off  the  gases  of  combustion,  not  to  produce  any  portion  of  the  head 
across  the  grates.  Thus,  when  the  stoking  doors  are  opened,  there  is 
no  tendency  to  draw  the  air  in  the  stoking  room  into  the  furnace; 
but,  on  the  other  hand,  there  may  be  a  tendency  for  the  heated  gases 
within  to  come  out  throiigh  the  open  doors,  as  observed  by  Mr.  Fether- 
ston   in   several  British  installations.     Forced   draft,  subject   to   close 


24  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Mr.  Venable.  regulation,  is  preferable  in  any  installation,  and  is  a  very  great  safe- 
guard against  the  admission  of  too  much  air  into  the  furnace,  above 
the  fires;  but  it  is  not  the  only  way  in  which  this  can  be  safeguarded, 
and,  in  many  plants,  especially  in  the  smaller  towns,  the  advantages 
to  be  derived  from  the  installation  of  a  boiler  are  not  as  great  as  the 
disadvantages. 

These  remarks  may  be  considered  as  not  properly  applying  as  a 
discussion  of  Mr.  Fetherston's  valuable  paper,  one  of  the  premises  of 
which  is  that  the  ashes  are  to  be  burned.  While  fully  recognizing 
this,  as  a  condition  precedent  to  his  inquiry,  and  having  no  quarrel 
with  it,  in  the  case  of  the  Borough  of  Richmond,  the  writer  has 
ventured  these  remarks  as  perhaps  of  some  interest  to  others,  for  con- 
ditions which  may  differ  from  those  stated  in  this  paper. 
Mr.  Gary.  ALBERT  A.  Cary,  Esq. — This  paper,  viewed  from  the  standpoint  of  a 
furnace  and  fuel  specialist,  is  of  great  interest  to  the  speaker,  who, 
having  had  considerable  experience  in  burning  various  fuels  of 
low  calorific  value  and  also  fuels  carrying  large  percentages  of 
moisture,  such  as  spent  tan  bark,  wet  refuse  wood-pulp  shavings,  spent 
licorice  root,  bagasse,  etc.,  can  well  appreciate  the  difiiculties  en- 
countered in  burning  wet  municipal  wastes;  and  burning  them  so  as 
to  obtain  sufficient  heat  for  steam-making,  which  heat  is  in  excess  of 
that  required  to  evaporate  the  moisture  contained  in  the  fuel,  and  to 
dissociate  the  fuel  (thereby  liberating  the  volatile  gases  it  contains, 
vdiich  action  is  necessary  before  these  fuel  constituents  can  burn). 

If  any  fuel  be  dried  and  an  analysis  be  made  of  its  chemical  com- 
position, then,  by  the  use  of  a  modification  of  the  well-known  Dulong 
formula,  a  determination  of  the  calorific  vak;e  of  the  sample  analyzed 
can  be  made.  This  may  or  may  not  be  of  use  in  furnace  determina- 
tions, depending  on  the  nature  of  the  fuel  and  the  value  of  the  sample 
as  a  fair  representative  of  the  entire  mass  of  fuel  consumed. 

It  is  no  easy  matter  to  obtain  a  representative  sample  of  the  entire 
fuel  consumed  during  a  test,  even  when  the  fuel  is  fairly  uniform  in 
quality,  but  when  its  quality  is  of  a  very  variable  nature,  such  as  in 
refuse-burning  plants,  the  difficulties  in  obtaining  a  small  sample  of 
fair  average  value  become  almost  insurmovmtable.  Aside  from  this 
difficulty,  after  making  calculations  from  the  chemical  analysis  of  a 
dried  sample,  one  does  not  obtain  a  true  calorific  value  of  the  fuel,  as 
this  process  of  determination  assumes  that  the  combustion  is  wholly 
an  exothermic  process,  that  is,  one  producing  heat  with  no  heat 
absorption  occurring  for  internal  or  external  reactions.  Such  endo- 
thermic  actions  always  take  place  in  the  process  of  combustion,  as  it 
requires  heat  energy  to  break  up  solid  masses  of  fuel  and  liberate  and 
split  up  the  hydro-carbons,  to  say  nothing  of  the  energy  required  to 
evaporate  the  moisture,  both  on  the  surface  and  contained  hydro- 
scopically. 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  25 

For  this  reason,  when  careful  tests  are  made,  the  calculated  fuel  Mr.  Cary. 
values  are  found  to  be  higher  than  those  obtained  by  using  the  oxygen 
fuel   calorimeter;    the   difference   between   these   two   values    indicates 
the  heat  energy  absorbed  by  endothermic  reactions. 

Coming  now  to  the  fuel  calorimeter,  reference  will  be  made  only 
to  the  work  done  in  the  Mahler  bomb.  By  a  proper  manipulation  of 
this  apparatus,  there  is  no  difficulty  in  determining  the  true  value  of 
the  sample  tested,  and  the  results  obtained  will  require  no  correc- 
tions for  chemical  endothermic  actions;  but  here,  also,  there  is  diffi- 
culty in  obtaining  correct  samples,  representing  a  fair  average  of  the 
whole  fuel  mass,  and  it  must  be  remembered  that  the  quantity  of 
fuel  tested  weighs  only  1  g.  (that  is,  less  than  0.04  lb.). 

The  great  difficulty  in  obtaining  the  true  calorific  values  of  the 
refuse  by  either  of  these  methods,  therefore,  can  be  well  appreciated, 
and  the  question  naturally  is:  How  can  this  most  important  value  be 
determined  ? 

The  answer  is  prompted  by  a  somewhat  extended  experience  in 
making  furnace  investigations  leading  to  accurate  heat  balances. 
Heat  balances  are  usually  obtained  by  calculations  made  from  the 
analysis  of  the  fuel,  the  quantity  of  fuel  used,  the  analysis  of  the 
products  of  combustion,  and  a  proper  consideration  of  the  various 
furnace  losses. 

By  a  somewhat  reversed  method  of  calculation,  made  from  a  series 
of  observations,  the  chemical  composition  of  the  fuel  may  be  obtained, 
and  from  it  the  unknown  quantity  to  be  determined,  then,  with  a  very 
fair  degree  of  accuracy,  its  value  may  be  found. 

The  accuracy  of  such  a  determination  is,  unquestionably,  far  greater 
than  may  be  obtained  by  any  system  of  sampling  when  such  a  mixed 
fuel  as  municipal  wastes  is  used. 

This  method  was  used  in  the  work  of  Mr.  C.  E.  Stromeyer,  re- 
ferred to  on  page  957,*  but  he  did  not  carry  his  work  far  enough  to 
obtain  sufficiently  satisfactory  results. 

The  furnace  gas  analysis  becomes  a  most  important  matter  in  such 
test  work,  and  the  mere  finding  of  the  percentages  of  COg,  O,  CO 
and  N,  by  difference,  by  the  use  of  the  ordinary  Orsat  apparatus,  will 
not  give  sufficient  information,  as  experience  has  taught  the  speaker 
that  in  such  work  it  is  necessary  to  determine  the  free  hydrogen  and 
hydro-carbons  as  well. 

Mr.  Stromeyer  also  relates,  in  his  report,  the  failure  of  his  high- 
temperature  measuring  apparatus,  which  furnishes  most  important 
information.  The  speaker  is  continually  using  such  apparatus,  with- 
out trouble,  in  furnace  tests  where  much  higher  temperatures  exist. 

The  speaker  does  not  wish  to  be  understood  as  criticizing  Mr. 
Stromeyer  in  this  work,  on  the  contrary,  he  regards  it  as  very  much 

*  Proceedings,  Am.  Soc.  C.  E.,  for  November,  1907. 


2G  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Mr.  Gary,  in  advance  of  any  testing  work  previously  done  in  refuse  crematories. 
He  merely  wishes  to  indicate  that  this  is  the  most  reliable  way  of 
obtaining  this  much  sought  for  information,  when  proper  testing  is 
done. 

To  obtain  data  needed  to  make  such  fuel  determinations,  one  does 
not  require  any  other  apparatus  than  that  used  in  making  complete 
and  exhaustive  furnace  tests,  but  careful  refinements  must  not  be 
neglected,  both  in  applying  and  using  the  apparatus  and  in  having 
them  all  carefully  calibrated. 

The  work  done  by  Mr.  Fetherston,  as  shown  in  this  paper,  to 
obtain  such  information,  by  fuel  sampling  methods,  is  certainly  highly 
creditable,  and  the  amount  of  work  involved  appeals  to  the  speaker 
strongly,  as  he  knows  by  experience  what  it  means. 

Concerning  the  large  percentages  of  moisture  held  in  fuels,  the 
speaker  has  profitably  passed  very  wet  fuel  between  a  pair  of  large 
cast-iron  rolls,  with  rough  faces,  one  roll  being  of  a  little  greater 
diameter  than  the  other.  These  rolls,  both  running  at  the  same  num- 
ber of  revolutions  per  minute,  were  held  together  by  large  springs 
which  allowed  them  to  separate  when  solid  chunks  reached  them. 

In  this  way  a  large  quantity  of  the  contained  moisture  was  squeezed 
out,  and  higher  temperatures  were  obtained  in  the  furnace,  as  well 
as  better  combustion,  for  the  furnace  is  the  poorest  place  in  the  world 
to  evaporate  water. 

Mr.  Fetherston's  statement  of  the  requirements  necessary  for  burn- 
ing wet  fuel,  on  page  972*  which,  as  he  states,  originated  with  Pro- 
fessor Thurston,  may  be  found  in  the  Journal  of  the  Franklin  Insti- 
tute for  1874,  where  it  will  be  found  to  refer  especially  to  spent  tan 
and  wet  saw-dust. 

For  the  combustion  of  moist  fuel,  the  highest  fvirnace  temperatures 
possible  are  most  essential,  and  that  requirement  is  one  of  the  weakest 
features  in  general  garbage  incinerating  plants.  It  is  firmly  believed 
that  much  profitable  development  is  possible  in  this  direction. 

The  disposition  of  the  highly  heated  surrounding  surfaces  men- 
tioned is  a  matter  of  much  importance  with  such  fuel,  and  combus- 
tion chambers  must  be  proportioned  to  the  amount  of  gaseous  matter 
and  moisture  given  off  by  the  fuel. 

The  speaker  can  hardly  admit  the  statement  indicating  that  com- 
bustion shoidd  be  retarded  and  limited  by  spots  of  dry  fviel  forming 
on  the  grate  and  burning  to  expose  wet  fuel,  thereby  stopping  com- 
bustion. Such  conditions  should  never  exist  as  they  indicate  bad 
design. 

To  obtain  the  most  desirable  results,  the  combustion  of  the  fuel 
should  constantly  be  accelerated. 

Pre-heated  air,  introduced  under  some  pressure,  to  offset  its  dilated 


*Proceedings,  Am.  Soc.  C.  E.,  for  November,  1907. 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  27 

condition,   will   assist   in  producing  such   results,   as   is   noted  by  Mr.  Mr.  Gary. 
Fetherston. 

Steam  jets  should  certainly  be  avoided  as  much  as  possible,  as  there 
is  altogether  too  nuich  steam  given  off  from  the  fuel  in  the  furnace, 
and  steam  has  a  cooling  effect  on  the  fire-bed.  To  obtain  the  best 
results,  the  steam  used  to  disintegrate  the  clinker  should  be  super- 
heated. 

The  speaker  cannot  agree  with  Mr.  Fetherston  when  he  places  the 
minimum  desirable  furnace  temperature  as  low  as  1  250°  fahr.,  which 
is  dangerously  near  the  lowest  temperature  at  which  some  of  the  gases 
found  in  the  furnace  will  ignite.  Such  a  temperature  will  surely  be 
followed  by  most  imperfect  combustion. 

A  furnace  should  not  fall  below  1  800°  fahr.,  as  experience  proves 
that,  under  lower  temperatures,  both  furnace  and  boiler  efficiencies 
drop.  Further,  2  000°  fahr.  is  too  low  for  a  maximum  temperature, 
as  the  speaker's  best  furnace  results  have  always  been  obtained  with 
temperatures  of  2  500°  fahr.,  or  greater. 

If  a  furnace  is  properly  designed  and  built,  there  is  no  reason  why 
it  should  not  be  durable  under  a  temperature  of  2  500°  fahr.,  and  with 
destructor  furnace  conditions. 

The  speaker's  experience,  of  many  years  in  furnace  work,  has 
taught  him  that  proper  provision  for  great  expansion  and  contraction 
is  frequently  neglected,  and,  also,  that  high-grade  refractory  materials 
are  not  used  as  much  as  they  should  be,  and  that  high-grade  furnace 
masons  are  not  employed,  but,  where  all  these  requirements  are  met, 
the  durability  of  furnaces  is  greatly  increased. 

On  page  970,*  it  is  noted  that  Mr.  Fetherston  assumes  a  combined 
furnace  and  boiler  efficiency  of  50  per  cent.  By  the  system  of  testing 
referred  to  in  this  discussion,  the  exact  efficiency  of  the  furnace  can 
be  obtained.  The  information  thus  obtained  will  also  point  out 
deiinitely  the  exact  causes  of  inefficiency,  and  thereby  lead  to  a  rapid, 
rational,  and  scientific  development  and  improvement  of  the  system  of 
garbage  incineration,  and  the  time  is  certainly  favorable  for  work  of 
this  nature,  as  shown  by  Mr.  Fetherston's  earnest  work  and  careful 
investigation  of  existing  conditions. 

E.  H.  Foster,  M.  Am.  Soc.  C.  E.   (by  letter). — The  valuable  data  Mr. Foster, 
which  Mr.   Fetherston  has  presented   in  this  paper  will  certainly  be 
appreciated    by    engineers    who    have    occasion    to    explore    this    com- 
paratively obscure  field,  and  it  is  certain  that  the  paper  will  prove  an 
important  addition  to  the  Society's  Transactions. 

Attention  is  called  to  the  quotation  on  page  972*  from  Professor 
Thurston,  giving  the  requirements  for  success  in  burning  wet  fuel: 
To  insure  that  "the  rapidity  of  combustion  may  be  precisely  equal  to 
*Proceedings,  Am.  Soc.  C.  E.,  for  November,  1907. 


28  DISCUSSION  ON  MUNICIPAL  EEFUSE  DISPOSAL  [Papers. 

Mr.  Foster,  and  never  exceed  the  rapidity  of  desiccation"  offers  a  condition  which 
is  to  be  steadfastly  striven  for,  but  which,  unfortunately,  can  only  be 
obtained  under  the  most  ideal  conditions,  and  one  can  only  hope  to 
fulfil  it  when  the  combined  collection  of  the  city's  waste  is  to  be 
burned.  So  long  as  furnaces  are  required  to  burn  garbage  only,  or 
garbage  and  rubbish,  special  provision  must  be  made  for  carrying  out 
the  above  requirements.  When  garbage  alone  is  burned,  fuel  must  be 
added  to  support  combustion.  Professor  Thurston's  remarks  show  why 
coal  should  be  used,  and  not  oil  or  natural  gas,  since  it  is  the  heated 
mass  in  the  coal,  and  not  the  volatile  matter,  which  accomplishes  the 
drying  process.  The  quality  of  coal  need  not  be  high;  in  fact,  coals 
of  the  poorer  quality,  containing  high  percentages  of  ash,  are  really 
more  suitable  for  this  purpose.  When  garbage  or  rubbish  have  to  be 
dealt  with,  without  ashes,  some  other  means  must  be  adopted  for  pre- 
paring the  garbage  for  burning,  and,  whatever  method  is  used,  it 
must  be  carried  out  inside  the  furnace,  thus  it  becomes  a  part  of  the 
detail  of  the  design. 

It  must  not  be  considered  that  the  chief  desideratum  is  to  obtain 
the  highest  possible  temperature  in  any  part  of  the  furnace.  Such  an 
impression  would  be  entirely  wrong.  The  temperature,  on  the  contrary, 
must  be  maintained  between  certain  moderate  limits,  preferably  be- 
tween 1  800  and  2  000°  fahr.,  but  with  a  minimum  never  less  than  1 250° 
fahr.     The  disadvantages  of  too  high  a  temperature  may  be  stated  as: 

Excessive  cost  of  repairs ; 

Melting  of  the  dust  and  clinker,  causing  it  to  fetick  to  the  fire- 
brick linings  inside,  and  loss  of  time  and  labor  in  cutting 
out,  periodically,  by  hammer  and  chisel,  the  slag-like 
accumulations ; 

Discomfort  to  the  operators  in  removing  the  clinker  from  the 
furnace. 

The  limit  of  low  temperature  is  reached  at  the  point  where  the 
gases  of  combustion  cease  to  be  dissociated  and  oxidized.  It  is  neces- 
sary, then,  to  maintain  a  temperature  well  above  this,  thus  rendering 
them  thoroughly  innocuous,  and  without  which  no  process  of  destruc- 
tion may  be  termed  sanitary  or  without  nuisance. 

In  the  study  of  designs  of  various  furnaces  intended  for  destroying 
refuse,  attention  has  been  drawn  to  the  conservation  of  the  heat  by 
various  recuperating  devices,  such  as  an  air  heater  for  extracting  the 
heat  from  the  flue  gases  and  transferring  it  to  the  air  which  is  being 
fed  to  the  furnace,  and  the  method  of  cooling  the  clinkers  by  taking 
up  their  heat  in  the  same  air  going  to  the  furnaces.  All  these  devices, 
which  resemble  the  conventional  economizers  and  air  heaters  used  in 
connection  with  steam-power  plants  for  securing  higher  efficiency  and 
economy   in  fuel  consumption,  serve  an  entirely  different  purpose  in 


Papers.]  DISCUSSION  ON  MUNICIPAL  EEFUSE  DISPOSAL  29 

the  case  of  destructor  furnaces.     They  are  rendered  necessary  in  order  Mr.  Fos^ter. 
to  insure  a  high  minimum  temperature  which  must  at  all  times  be 
maintained,  the  character  of  the  material  fed  to  the  grates  being  of 
such  a  nature  that  these  precautions  are  necessary. 

Whereas  the  steam  generated  from  the  plant  represents  a  valuable 
asset,  and,  in  some  instances,  can  be  made  to  do  useful  work,  doubtless 
there  will  be  many  cases  where,  in  the  absence  of  a  suitable  method  of 
utilizing  this  power,  the  steam  must  be  blown  off  and  wasted.  As  it 
is  the  condition  in  the  furnace  which  is  of  most  importance,  even  if 
the  steam  generated  is  wasted,  the  devices  for  recovering  the  heat  must 
be  used. 

On  page  978*  Mr.  Fetherston  siiggests  a  further  improvement 
which  might  be  made  in  the  Westmount  plant,  namely,  to  utilize  the 
heat  contained  in  the  hot  clinker  for  raising  the  temperature  of  the 
air  for  combustion.  This  idea  is  being  carried  out  in  the  city  plant 
now  under  construction. 

It  is  a  mistake  to  rely  on  the  recommendation  that  the  conversion 
of  the  power  of  the  destructor  plant  into  electrical  energy  is  the  most 
suitable  outlet  for  that  power.  Mr.  Fetherston  mentions  the  pumping 
of  water  or  sewage  as  an  appropriate  use.  A  still  inore  appropriate 
use  would  seem  to  be  the  manufacture  of  ice,  for  which  such  a  plant 
is  strikingly  well  adapted.  For  instance,  with  an  absorption  ice  ma- 
chine, 9  lb.  of  ice  may  be  readily  procured  by  the  burning  of  1  lb.  of 
coal  imder  the  grates  of  the  boiler,  whereas,  in  a  destructor  plant  of 
50  tons  capacity,  1  lb.  of  steam  may  be  readily  evaporated  per  pound 
of  mixed  refuse  destroyed.  A  50-ton  destructor  plant  would  serve  a 
community  with  a  population  of  approximately  40  000.  By  comparing 
these  figures  it  will  be  seen  that  a  50-ton  refuse  destructor  will  pro- 
duce 50  tons  of  ice  per  day,  or  an  allowance  of  2J  lb.  of  ice  per 
capita,  which  would  be  a  liberal  amount. 

An  important  feature  in  the  design  of  furnaces  is  the  avoidance 
of  smoke;  this  can  only  be  accomplished  by  isolating  completely  the 
fiirnace  and  combustion  chamber  from  the  water-heating  surface  con- 
nected with  the  boiler,  as  the  chilling  effect  produced  by  contact  of  the 
partially  cooled  consumed  gases  against  the  cold  surface  of  the  tube 
containing  water  will  suppress  complete  combustion  and  result  in  a 
smoky  chimney. 

B.  F.  Welton,  Assoc.  M.  Am.  See.  C.  E.— The  speaker  has  fol-  ^r.  Weiion. 
lowed  with  great  interest  the  development  of  the  work  done  by  Mr. 
Fetherston,   the  results   of   which   are   so   admirably   presented   in  his 
valuable  paper. 

The  speaker's  relation  to  this  work,  as  stated  by  the   author,  has 
been  in  connection  with  the  determination  of  the  calorific  values  in- 
corporated in  the  text  of  the  paper.     The  purpose  of  this  discussion, 
*  Proceedings,  Am.  Soc.  C.  E.,  for  November,  1907. 


30  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Mr.  Welton.  therefore,  is  to  describe  in  some  detail  the  methods  used  in  making 
the  calorific  tests  and  proximate  analyses,  in  order  that  the  reader  may 
be  enabled  to  form  his  own  estimate  of  the  relative  value  of  those  re- 
sults as  compared  with  similar  tests  of  other  and  more  homogeneous 
materials  used  as  fuel. 

It  is  also  desired  to  record  the  results  of  a  series  of  chemical 
analyses  of  the  component  parts  of  the  refuse  made  in  the  same 
laboratories  by  Professor  Stephen  F.  Peckham,  Member  of  the  Ameri- 
can Chemical  Society,  whose  assistance  has  been  highly  appreciated. 

The  primary  purpose  of  the  experiments  was  to  provide  funda- 
mental data  from  which  could  be  determined  the  feasibility  of  the 
sanitary  disposal  of  the  wastes  of  the  Borough  of  Richmond  by  self- 
combustion,  in  a  refuse  destructor  of  the  same  general  type  as  used 
in  Great  Britain. 

Inasmuch  as  the  matter  of  heat  utilization  and  power  production 
was  to  be  taken  up  ultimately,  in  connection  with  the  disposal  of  the 
refuse,  the  resialts  of  the  experiments  were  also  to  be  considered  as 
possibly  affecting  the  design  of  the  destructor.  If  the  results  of  the 
calorific  tests  should  show  that  the  material  was  not  suitable  for  self- 
incineration,  it  was  hoped  that  the  chemical  analyses  might  provide 
the  necessary  information  for  determining  some  alternative  method 
of  sanitary  disposal.  On  the  other  hand,  if  the  material  should  be 
found  suitable,  the  chemical  analyses  might  furnish  additional  data 
for  the  stvidy  of  means  for  the  prevention  of  possible  nuisance  by  the 
escape  of  the  products  of  combiistion,  or  for  tlie  recovery  of  com- 
mercially valuable  material. 

After  the  conclusion  of  trial  calorific  tests  on  two  sets  of  samples 
to  ascertain  what  methods  of  handling  the  material  in  the  laboratory 
would  secure  the  desired  uniformity  of  results,  a  consultation  was  held 
between  the  author  and  the  speaker  to  define  the  scope  of  the  experi- 
ments. 

It  was  decided : 

First. — That,  if  the  experiments  were  to  be  conclusive,  they  should 
be  extended  over  a  period  of  at  least  a  full  year,  thus  showing  the 
entire  seasonal  variation  in  the  character  of  the  collections,  which 
variation,  it  was  thought,  might  be  sufficient  to  interfere,  perhaps,  to 
a  serious  degree,  with  the  successful  operation  of  a  destructor; 

Second. — That  the  samples  should  be  taken  with  sufficient  fre- 
quency and  in  such  manner  as  to  be  truly  representative  of  the  col- 
lections, both  as  regards  the  character  of  the  material  and  the  period 
covered. 

It  was  finally  settled  tliot  the  samples  submitted  to  the  laboratory 
should  represent  the  daily  collections  of  the  Bureau  of  Street  Cleaning 
for  a  period  of  about  two  weeks,  or  a  half  month. 


Pitpcrs.]  DISCL'SSION   ()\    MLXiCirAL   KEL^USE  DISi'OSAL  31 

The  primary  sampling-  from  the  actual  collections,  as  well  as  the  Mr.  Weltoc. 
initial  preparation  of  the  half-monthly  samples,  was  to  be  made  under 
the  direction  of  Mr.  Fetherston. 

The  sampling,  as  described  in  detail  by  the  author,  consisted  in  the 
selection  of  representative  material  which  was  subsequently  separated, 
by  sieves  and  hand-picking,  into  six  general  classes  as  follows : 

1. — Garbage, 

2. — Coal  and  cinders, 

3.— Rubbish, 

4. — Fine  ash, 

5. — Clinker, 

6. — Incombustible  material. 

The  garbage  consisted  of  vegetable  and  animal  matter,  etc.,  such 
as  ordinarily  collected  from  dwellings. 

The  coal  and  cinders  was  the  better  portion  of  the  stove  and  furnace 
wastes  of  the  district. 

The  rubbish  consisted  of  a  variety  of  materials,  such  as  paper, 
excelsior,  rags,  iibrous  material,  etc. 

The  fine  ash  was  the  material  from  the  general  collections  which 
would  pass  through  a  screen  of  f-in.  mesh,  and  consisted  principally 
of  the  finer  residue  from  domestic  fires. 

The  clinker  was  that  contained  in  the  residue  from  domestic  fires, 
and  those  of  schools,  churches,  etc. 

The  incombustible  material  was  largely  glass,  metal,  stone,  bricks, 
etc. 

The  initial  preparation  of  the  samples  comprised  the  reduction  of 
large  quantities  of  material  of  the  several  classes  by  quartering,  the 
evaporation  of  nearly  all  the  moisture,  and  the  rough  pulverizing  of 
all  samples  to  effect  a  uniformity  which  would  serve  to  make  the 
samples  submitted  to  the  laboratory  truly  representative. 

The  weight  of  these  samples  was  approximately^: 

1  lb.  of  garbage  (dry)  ; 

2  lb.  each  of  coal  and  cinders,  clinker  and  fine  ash; 
5  lb.  of  rubbish. 

The  condition  of  the  samples,  as  they  arrived  at  the  laboratory, 
after  going  through  this  preliminary  process,  was  about  as  follows : 
The  garbage,  in  the  majority  of  cases,  was  fairly  dry,  but  soft  and 
greasy ;  most  of  it  would  pass  a  sieve  of  ^-in.  mesh,  and,  while  the 
odor  was  decidedly  in  evidence  it  was  not  offensively  so.  Nearly  all 
the  coal  and  cinders  would  pass  a  sieve  of  I -in.  mesh,  and  showed  a 
large  proportion  of  unburned  coal. 

The  fine  ash  and  clinker  were  in  about  the  same  condition  as  the 
coal  and  cinders,  except  that  the  difference  in  the  quantity  of  carbon 
present  was  plainly  evident  from  the  color  and  general  appearance. 


32  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Mr.  AVelton.         The  rubbish  presented  the  appearance  of  shredded  rags,  paper,  etc. 

No  incombustible  material  was  tested,  for  obvious  reasons. 

Upon  arrival  at  the  laboratory,  each  sample  was  immediately 
placed  in  a  wide-mouthed  glass  jar  with  a  ground-glass  stopper,  and 
as  soon  as  convenieitt  thereafter  a  careful  determination  was  made  of 
the  contained  amount  of  moisture.  This  operation  was  conducted 
using  about  10  g.  of  garbage  and  about  5  g.  each  of  the  other  samples. 

The  whole  of  each  sample  of  garbage  and  rubbish  was  then  made 
to  pass  a  sieve  of  No.  20  mesh  by  repeated  grinding  in  a  small  pul- 
verizer of  the  coffee-mill  type.  The  coal  and  cinders  sample  was 
pulverized  in  a  laboratory  ball  mill  until  it  would  all  pass  a  No.  40 
sieve.     The  clinker  and  fine  ash  were  treated  in  the  same  manner. 

The  samples  were  then  replaced  in  their  respective  glass  jars  and 
thoroughly  mixed  by  agitation. 

Proximate  analyses  were  next  made,  determining  again  the 
moisture,  and,  in  addition,  the  volatile  matter,  fi.xed  carbon,  and  ash. 
For  these  determinations,  the  following  weights  of  material  were  used: 

Garbage    about  1.5  g. 

Coal    and   cinders "       2.0  " 

Clinker "       2.0  " 

Fine   ash    "       2.0  " 

Kubbish    "      0.5  " 

These  quantities  of  the  several  materials  were  taken  at  random 
directly  from  the  jar  containing  the  whole  sample,  since  it  was  found 
that  practical  duplication  of  results  could  readily  be  obtained  without 
further  reduction  in  size  or  quantity  of  the  sample. 

All  determinations  of  moisture  were  made  by  using  an  electric 
oven  kept  at  a  constant  temperature  of  about  180°  fahr.  The  coarse 
samples  were  allowed  to  remain  in  the  oven  for  about  18  hours,  but 
only  about  1^  hours  were  necessary  when  the  samples  were  in  the 
pulverized  condition.  The  volatile  matter  was  determined  by  placing 
the  dried  material  in  a  small,  covered,  porcelain  crucible,  over  a  three- 
flame  Bunsen  burner,  care  being  taken  that  all  the  carbon  deposited 
during  the  combustion  of  the  volatile  hydro-carbon  was  afterward  con- 
sumed by  the  Bunsen  flame.  (Platinum  crucibles  were  first  used  for 
this  work,  but  some  constituent  of  the  coal  and  cinders,  which  was 
later  discovered  to  be  tin,  probably  from  tin  cans,  re-acted  with  the 
platinum,  ruining  the  crucibles,  and  their  use  was  abandoned). 

The  fixed  carbon  then  remaining  in  the  crucible  was  next  reduced 
to  ash  by  open  burning  over  the  same  Bunsen  flame  until  no  loss  in 
weight  occurred. 

The  percentages  of  the  various  determinations  were  reduced  by 
calculation  to  the  basis  of  the  condition  of  samples  as  they  were  re- 
ceived at  the  laboratory,  and  the  garbage  analyses  were  still  further 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  33 

modified   to   represent  the   conditions   in   the  original  sample   before  Mr.  Weitou. 
evaporation  of  any  of  its  moisture,  a  record  of  the  evaporative  tests 
being  sent  to  the  laboratory  with  each  sample. 

The  calorific  values  were  determined  by  the  Mahler  bomb  calorim- 
eter, which  provides  for  the  combustion  of  the  material  in  the  presence 
of  oxygen  at  a  pressure  of  25  atmospheres. 

The  tests  were  made  using  the  following  weights  of  material : 

Garbage    0.80  to  1.00  g. 

Coal  and  cinders 0.50  «  0.Y5  " 

Kubbish    0.35  "  0.50  " 

Fine  ash  0.50  "  0.75  " 

Clinker    0.50  "  0.75  " 

These  values,  as  obtained  by  the  actual  tests,  were  reduced  to  values 
per  pound  of  dry  sample,  original  sample,  and  combustible,  in  that 
order,  using  the  corrected  proximate  analyses  as  a  basis.  These  are 
the  figures  that  appear  in  Tables  6  and  7. 

There  was  no  diflSculty  in  securing  satisfactory  combustion,  ex- 
cept in  the  tests  of  "fine  ash"  and  "clinker,"  in  which  the  percentage 
of  inert  matter  was  so  high  that  it  prevented  ignition  of  the  com- 
bustible portion  of  the  sample  by  the  ordinary  means.  In  these  cases, 
therefore,  a  small  amount  of  naphthaline  was  introduced  with  the 
sample  to  start  the  combustion,  and  a  deduction,  representing  the 
calorific  value  of  the  naphthaline  used,  was  made  subsequently. 

The  residue  from  the  combustion  of  the  garbage  was  hard,  vitreous, 
and  invariably  in  the  form  of  small  globules  of  a  brownish  black  color. 
That  of  the  coal  and  cinders  was  naturally  about  the  same  in  appear- 
ance as  the  ash  of  anthracite  coal,  while  the  rubbish  left  little  more 
that  could  be  seen  with  the  eye  than  a  stain  on  the  combustion  tray. 

At  the  beginning  of  the  experiments,  tests  were  made  in  duplicate 
on  all  samples  until  it  became  evident  that  the  differences  in  results, 
as  shown  by  the  duplications,  were  well  within  the  variation  that 
might  easily  occur  in  the  primary  selection  of  representative  samples. 
In  amount,  these  differences  were  generally  less  than  1%  of  the 
calorific  value  of  the  dry  material.  By  this  time,  also,  the  uniformity 
in  the  character  of  each  class  of  material,  as  shown  by  the  calorific 
value  per  pound  of  combustible,  began  to  be  noticeable,  and  it  was 
observed  that  this  value  would  serve  to  detect  errors  in  manipulation 
and  computation  as  well  as  to  indicate  the  occasions  when  duplication 
was  required.  As  a  consequence,  tests  on  the  same  sample  were  rarely 
repeated  thereafter,  unless  the  value  per  pound  of  combustible  was  at 
some  variance  with  the  average  of  the  other  tests  already  made. 

To  those  who  are  not  familiar  with  the  calorific  values  of  the 
staple  fuels,  such  as  anthracite  and  bituminous  coals,  it  may  appear 
that  no  great  confidence  should  be  placed  in  the  results  of  these  tests 


34  DISCUSSION  ON  MUNICIPAL  KEFUSE  DISPOSAL  [Papers. 

Mr.  Weiton.  on  material  which  would  naturally  be  expected  to  vary  widely  in 
character.  As  a  matter  of  fact,  the  experiments  have  shown  a  uni- 
formity of  character  in  the  material  which  is  all  the  more  remarkable 
in  that  it  was  not  anticipated.  Indeed,  now,  when  all  the  data  are  at 
hand,  the  conclusion  might  easily  be  drawn  that  in  the  instances  where 
the  largest  variations  in  calorific  values  per  pound  of  combustible 
occur,  this  variation  is  more  likely  to  be  due  to  the  difficulty  of  ob- 
taining representative  samples  from  the  collections  than  from  actual 
differences  in  character. 

Moreover^  few  who  have  had  no  occasion  to  study  the  matter  of 
analyses  and  calorific  tests  of  coal  are  aware  of  the  variation  in  fuel 
value  of  its  combustible  portion  or  what  is  known  as  "pure  coal." 

In  this  respect  the  figures  in  Table  12  are  of  interest.  These 
are  deduced  from : 

First. — The  report  of  the  coal-testing  plant  of  the  United  States 
Geological  Survey  at  St.  Louis,  in  1904; 

Second.^ — From  records  of  the  Department  of  Water  Supply, 
Gas  and  Electricity,  at  Mount  Prospect  Laboratory,  New 
York  City. 

The  chemical  analyses  made  by  Professor  Peckham  consisted  of 
organic  analyses  of  composite  samples  representing  the  collections  of 
the  entire  period,  and  inorganic  analyses  of  the  residue  from  burning 
the  same  over  a  Bunsen  flame.  They  will  not  be  described  in  detail 
here,  but  the  results  of  both  series  of  analyses  have  been  combined  in 
Table  13. 

These  results  would  have  been  included  in  Mr.  Fetherston's  paper, 
except  for  the  fact  that  their  completion  was  delayed  by  pressure  of 
more  important  matters  in  the  laboratory,  and  they  have  only  very  re- 
cently become  available. 

The  same  reason  also  accounts  for  the  relatively  large  percentage 
of  undetermined  constituents  in  the  garbage  sample. 

Table  13  also  shows  a  comparison  between  the  calorific  value  of  the 
samples,  as  calculated  from  the  chemical  analyses,  and  as  determined 
by  the  calorimeter.  The  correspondence  is  believed  to  be  sufficiently 
close  to  serve  as  a  general  verification  of  the  entire  work.  The  per- 
centages in  Table  13  are  all  computed  on  the  weight  of  dry  samples 
as  a  basis.  The  presence,  in  considerable  amounts,  of  volatile  hydro- 
carbons in  the  garbage  and  rubbish  samples  may  be  noted  if  the  car- 
bon, as  shown  by  the  chemical  analyses,  be  compared  with  the  fixed 
carbon  of  the  proximate  analyses  reduced  to  a  dry-sample  basis. 

In  the  determination  of  the  moisture  in  the  garbage  samples  for 
the  proximate  analyses,  it  is  extremely  probable  that  some  of  the 
lighter  and  more  easily  volatile  hydro-carbons  were  driven  off  and 
computed  as  moisture.     This  is  undoubtedly  the  reason  why  there  is 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL 


35 


1 

o 

O 

A'1  P 

O  1  B 

!  p? 

3  o  c 

3  ^  SJ^ 

(I  »  So" 

O  t-n 

'-t 

1s= 

V  c 

>d  a 

o  S 

o  &" 

1 

c  * 

C  CD 

1 
I 

&g 

§-s 

0  o 

o  9 

^^^^    : 

■^> — 

:^ 

> 

o 

j_^ 

Range,  in 

cnn> 

2 

CO 

o 

British  thermal 

a 

3 

3d 
o 

units. 

o-.p 

its 

KisrO!^ 

>■ 

o 

o 

o 

o 
b 

Percentage 
of  average 

=\ 

=^ 

value. 

*:=!!<  Qa 

!>» 

8S 

t-i 

(» 

o 

Range,  in 

9^ 

4^ 

00 

British  thermal 

g 

05 

units. 

d 

10.0 

Pijrcentage 
of  average 

•§5! 

CO 

O 

G 

=^ 

^ 

value. 

^» 

o 

g 

w 

o 

3 
o 

HH 

o' 

d 

m 

CO 

Range,  in 

"—  2  5  S' 

p 

OQ 

00 

British  thermal 
units. 

CO 

i 

^5! 

o 

o 

'o 

00 

Percentage 
of  average 

=\ 

.^ 

value. 

Range,  in 

1 

50 

British  thermal 
units. 

il'wg 
33§°- 

r< 

"O  o  5  « 

E» 

W 

Percentage 

o3 

g- 

Vi 

of  averae-e 

s 

=\ 

^           , 

value. 

o 

g 

ts 

Range,  in 

^         Q 

Q 

a 

to 

British  thermal 

J§w2?S 

o 
3 
3 

^ 

o? 

§ 

units. 

oj  o  •-!  cr 

F 

S5  B-O  ?3 

W 

H 

3  a  CW3 

2  Pots  » 

M 

o» 

Percentage 
of  average 

'ESp'-'h* 

P 
CD 

a 

^ 

^ 

value. 

3 

O 

Ms 

1 

> 

s2 

Range,  in 

WW^ 

o 

Q 

OS 

British  thermal 

o 

a 

g 

units. 

P^o  S 

P  C  (B 

a 

«  ^Ha 

.fe 

s 

Percentage 

a  -  n 

i=.o  o 

S 

of  average 

■  "=3 

=^ 

^ 

value. 

Mr.  WelioB 


o 

M 
O 

O 

> 


2 
o 

<l 

t> 
r 

a 

CO 

B 

W 

>    ts 

M 

!>  CO 
f  H 
C)     CO 

H  "* 
CO     >■ 

■       !^ 
O 

H 

M 

H 

» 
O 

o 

w 

H 
O 
!z! 

C! 
O 
K 

W 

> 

O 

H 


bd 


36 


DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL 


[Papers. 


Mr.  Weitou.  not  a  closer  agreement  between  the  calorific  values  of  dry  samples  of 
garbage,  as  calculated  from  the  analyses,  and  as  determined  in  the 
calorimeter.  The  oxygen  is  determined  by  difference,  and  the  ash  is 
the  average  of  the  proximate  analyses  reduced  to  the  dry-sample  basis. 

TABLE    13. — Chemical    Analyses   of    Dry    Composite    Samples   op 
Coal  and  Cinders,  Garbage,  and  Kubbish,  Eepresenting  Col- 
lections FOR  the  Year  1905-06. 


Constituents. 


Percentage  by  weight  of: 

Carbon 

Hydrogen 

Nitrogen 

Oxygen 

Silica 

Iron  oxide  and  alumina 

Lime 

Magnesia 

Phosphoric  acid 

Carbonic  acid 

Lead 

Tin 

Alkalies  and  undetermined 


Coal  and 
cinders. 

Garbage. 

Rubbish. 

55.77 

43.10 

42.39 

0.75 

6.24 

5.96 

0.64 

3.70 

3.41 

2.37 

27.74 

33.52 

30.01 

7.56 

6.49 

8.98 

0.41 

2.03 

1.31 

4.26 

2.26 

Trace. 

0.28 

0.57 

None. 

1.47 

0.10 

None. 

0.59 

1.49 

Trace. 
Trace. 

1  Sulphides,  0.20  ] 

0.52 
Trace. 

0.27 

4.45 
100.00 

1.21 

100.00 

100.00 

Calorific  Values,  in  British  Thermal  Units. 


Calculated  from  above  analyses 

Average  of  calorimeter  determinations. . 


8  3S2 
8  510 


7  970 

8  351 


7  250 
7  251 


The  calculations  from  the  cliemical  analyses  are  made  as  follows : 

62  100  (h  —  ^)  +  14  500  C.  =  British  thermal  units. 
^         8 ' 

Mr.  Koyi.  C.  Herschel  Koyl,  Esq.  (by  letter). — This  paper  presents  in  ad- 
mirable form  the  results  of  a  very  careful,  systematic,  and  thorough 
study  of  the  possibility  of  destroying  by  fire  the  mixed  wastes  of  the 
Borough  of  Richmond  in  a  manner  innocuous,  inoffensive,  and  not  too 
costly. 

The  need  of  such  an  investigation  was  pressing,  and  its  value  not 
merely  local,  because  the  number  of  small  communities  in  America, 
in  which  this  problem  is  of  first  importance,  is  considerable  and 
growing. 

The  technical  question  is  whether  the  mixed  waste  contains  enough 
combustible  to  be  self-burning,  at  a  temperature  sufiiciently  high  to 
destroy  and  not  merely  distil  the  volatile  organic  matter.  Records 
show  that,  in  England  and  on  the  Continent,  a  satisfactory  disposal 
of  mixed  municipal  refuse  is  made  in  this  way.  but  it  is  also  known 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  37 

that  abroad  there  is  less  waste  of  edible  matter  than  in  the  United  Mr.  Koyi. 
States;  and,  therefore,  before  risking  $60  000  of  municipal  money,  it 
was  the  part  of  wisdom  to  determine  the  theoretical  fuel  efficiency  of 
the  waste  of  Staten  Island.  From  the  character  of  the  examination 
and  its  completeness,  Mr.  Fetherston's  results  may  be  accepted  with 
confidence,  and  also  his  conclusion  that  a  destructor  of  the  English 
type  will  burn  the  mixed  waste  of  the  Borough  of  Richmond  effec- 
tively. True,  the  expense  will  not  be  small;  but  if  the  destruction  of 
organic  matter  is  complete  and  inoffensive  to  the  neighborhood,  a  cost 
of  from  $1.00  to  $1.50  per  ton  should  not  be  prohibitive,  in  view  of 
the  fact  that  any  other  method  of  disposition  would  be  extremely 
difficult  in  Staten  Island. 

Regarding  the  limits  of  usefulness  of  these  waste  destructors:  It 
has  been  proven  by  eight  months'  operation  in  Westmount,  Montreal, 
that  an  average  of  20  tons  of  mixed  waste  per  day  can  be  destroyed  at 
a  working  cost  of  31  cents  per  ton,  and  a  total  cost  of  80  cents  per 
ton ;  therefore,  a  population  of  13  000  people  is  not  too  small  to  have 
the  economical  service  of  a  destructor. 

An  upper  limit,  however,  is  reached  when  considering  a  city  from 
which  there  is  enough  garbage  to  make  profitable  a  modern  reduction 
plant  to  separate  the  organic  matter  into  grease  and  fertilizer.  For 
instance,  the  City  of  New  York  makes  satisfactory  disposal  of  ap- 
proximately 3  000  000  tons  of  mixed  waste  per  year  (70%  ashes,  12% 
street  sweepings,  12%  garbage  and  6%  light  refuse,  by  weight)  at  an 
average  cost  of  40  cents  per  ton.  It  would  be  folly  to  talk  of  putting 
all  this  material  through  destructors  at  a  cost  of  75  cents  per  ton. 

Note  should  be  made  of  a  fact  not  mentioned  by  Mr.  Fetherston, 
that  in  the  "coal  and  cinders"  which  makes  27%  of  his  total  collec- 
tion, or  about  35%  of  his  ash  collection,  more  than  half  is  not  only 
burnable  coal,  but  salable  coal.  This  arises  from  the  fact  that  most 
of  the  coal  which  gets  into  the  ash-pit  is  imdersized  for  the  grate  and 
falls  through  unburned  and  indeed  unmarked  by  the  fire.  The  writer 
has  taken  from  many  sample  tons  of  Manhattan  ashes  an  average  of 
20%  of  salable  coal,  from  furnace  size  down,  of  which  about  half, 
after  being  washed,  was  indistinguishable  from  coal  fresh  from  the 
mine.  This  means  nothing  in  the  Borough  of  Richmond,  but  it  will 
be  the  determining  factor  in  settling  the  method  of  final  disposal 
where  anthracite  is  used  in  cities  which,  at  the  same  time,  are  large 
enough  to  make  profitable  the  mechanical  separation  of  the  coal  from 
the  clinker  and  ashes.  The  process  is  not  more  difficult  than  the  con- 
centration of  ore,  and  there  is  an  average  profit  of  about  $2  per  ton 
of  recovered  coal. 

In  the  United  States  the  advocates  of  reduction  "utilize"  garbage 
by  separating  it  into  water,  grease,  and  fertilizer.  The  advocates  of 
incineration  "utilize"  dry  refuse  by  picking  out  its  30%  of  salable 


38  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Mr.  Koyi.  paper,  rubber,  etc.,  before  they  burn  the  remaining  70%  of  rubbish. 
And  "utilization"  is  the  keynote  to  successful  policy  in  any  large  city. 
It  now  costs  New  York  $1250  000  for  the  final  disposition  of  its 
municipal  wastes.  It  would  cost  $2  250  000  to  put  all  the  waste 
through  destructors.  It  would  cost  about  $200  000  to  do  it  scientifi- 
cally and  save  what  ought  not  to  be  burned  or  buried,  as  follows: 

Profit  from  utili- 
zation of  rubbish 
Cost.  and  coal. 

Garbage    360  000  tons  (contract)      $200  000 

Street  sweepings 360  000      "     at  40  cents        144  000 

Ash  and  clinker.  ...    1  680  000      "     at  40  cents        672  000 
Kubbish   180  000      "     burned  at 

profit $40  000 

Coal 420  000     "    recovered 

at  profit 840  000 


$1  016  000         $880  000 
Less 880  000 


$136  000 

or,  the  Department  of  Final  Disposition  would  be  almost  self-support- 
ing. 

Mr.  Ti-ibus.  Louis  L.  Tribus,  M.  Am.  Soc.  C.  E. — This  paper  gives  evidence  of 
a  great  deal  of  work,  and  the  speaker  can  say  from  personal  knowledge 
that,  in  the  Borough  of  Richmond,  and  along  the  lines  described, 
there  has  been  a  vast  amount  of  work  which  does  not  appear  in  the 
paper,  yet  its  results  will  certainly  secure  great  advancement  in  the 
art  of  refuse  disposal. 

Prior  to  the  inauguration  of  the  Greater  City  of  New  York,  Staten 
Island  (then  becoming  the  Borough  of  Richmond)  was  occupied  by  a 
number  of  corporate  villages  and  a  great  many  small  hamlets,  the  latter 
controlled  by  the  usual  "township"  and  "county"  system  of  govern- 
ment, the  incorporated  portions  by  "village"  form,  with  more  or  less 
intelligent  management,  as  politics  determined. 

During  the  first  four  years  following  consolidation,  little  was  done, 
other  than  to  get  accustomed  to  being  a  part  of  the  great  city. 

On  January  1st,  1902,  under  the  revised  charter,  considerable  home 
rule,  and  a  borough  president  of  character  and  ability,  the  first  ad- 
vance toward  real  progress  was  made. 

Street  cleaning  and  refuse  disposal  had  been  cared  for,  to  a  very 
limited  extent,  during  the  preceding  four  years,  with  a  small  force 
of  men,  but  supervised  by  a  man  trained  under  Colonel  Waring,  who 
had  the  welfare  of  his  subject  at  heart.  The  speaker  was  called  upon 
by  the  borough  president  early  in  1902  to  act  in  both  professional  and 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  39 

executive  capacities,  and  take  charge  of  the  public  works  and  main- Mr.  Tribus. 
tenance  bureaus  of  the  borough.  He  was  given  very  free  rein  in 
securing  betterments  in  plan  and  operation,  but,  on  taking  charge,  in- 
sisted that  he  was  to  be  free  from  politics  in  the  work  itself.  Richard 
T.  Fox,  formerly  in  charge  of  the  work  in  Richmond,  as  noted  before, 
for  the  Department  of  Street  Cleaning  for  the  whole  city,  was  placed 
in  charge  of 'the  newly  created  "Bureau  of  Street  Cleaning,"  which 
at  this  time  came  under  the  President  of  the  Borough  through  the 
Commissioner  of  Public  Works.  Fairly  liberal  appropriations  were 
made,  so  that,  after  careful  plans  had  been  laid,  improvement  became 
the  order  of  the  day.  The  first  step  noted  was  the  banishment  of 
garbage  cans  and  refuse  receptacles  from  the  sidewalks  and  streets  as 
far  as  possible,  all  such  being  removed  by  collectors  from  behind  the 
buildings,  the  empty  cans  being  then  returned  to  their  places.  The 
next  step  toward  efficiency  and  the  establishment  of  esprit  de  forps 
among  the  men,  was  made  by  placing  them  all  in  uniform;  the  third 
step  was  to  employ  the  men  continuously  throughout  the  year,  so  as 
to  render  service  daily  instead  of  spasmodically.  This,  of  course,  ap- 
plied more  to  street  cleaning,  pure  and  simple,  than  to  refuse  collec- 
tion which,  formerly  also,  had  to  be  more  or  less  regular  throughout 
the  year. 

After  some  two  years'  service,  Mr.  Fox  accepted  a  call  to  the  City 
of  Chicago,  to  show  there  what  scientific  and  business  methods  could 
accomplish  in  the  way  of  street  cleaning  and  refuse  disposal. 

Mr.  Fetherston,  as  a  member  of  the  borough  engineer  corps,  had 
been  assigned  to  specific  work  in  connection  with  local  scientific  tests 
in  refuse  disposal,  in  which  work  he  gave  so  good  an  account  of  him- 
self that  when  Mr.  Fox  resigned  he  was  selected  to  take  the  place  of 
"Superintendent." 

The  paper  describes  very  clearly  the  course  taken,  which  has  led 
to  the  recommendation  and  the  actual  construction  of  the  first  refuse 
destructor  of  this  type  in  the  United  States.  It  is  confidently  hoped 
that  in  a  few  years  this  paper  will  be  supplemented  by  one  describing 
the  destructor  and  telling  of  its  efficient  operation.  That,  however, 
will  depend  largely  on  the  intelligence  exhibited  in  its  management, 
for  the  best  piece  of  machinery  may  give  poor  results  unless  well 
handled. 

In  studying  the  refuse  disposal  question  in  the  Borough  of  Rich- 
mond, it  has  been  necessary  to  estimate  very  carefully  the  probable 
development  of  this  specific  locality,  as  its  conditions  are  changing 
very  rapidly  year  by  year.  It  is  not  improbable  that,  within  the  life 
of  the  present  generation,  the  whole  island  will  be  practically  built 
up  with  residences,  factories,  stores,  and  valuable  water-front  improve- 
ments. This  would  mean  that  it  would  be  impossible  to  find  places 
for  the  burial  of  garbage,  for  maintaining  ash  dumps,  and  for  any 


40  DISCUSSION  ON  MUNICIPAL  KEFUSE  DISPOSAL  [Papers. 

Mr.  Tribus.  of  the  nuisance-producing  plants  for  the  destruction  of  garbage  by 
low-temperature  cooking.  All  experiments,  therefore,  in  the  past  six 
years  have  been  directed  toward  finding  a  process  that  would  convert 
refuse,  without  nuisance,  into  some  useful  or  innocuous  material.  The 
investigations  which  have  been  made  so  carefully  by  Mr.  Fetherston 
and  others  assigned  to  the  work  from  time  to  time,  therefore,  have 
been  directed  specially  to  this  system;  as,  by  process  of  elimination, 
all  other  systems  were  dismissed  as  not  suitable  for  the  probable  local 
conditions  of  development,  hence  the  conclusion  that  mixed  refuse 
destruction  promised  more  to  the  Borough  of  Richmond  than  did  any 
other  process;  though  it  should  be  clearly  understood  that  other  sys- 
tems might  be  more  advantageous  in  other  communities  under  differ- 
ent conditions,  and  time  may  prove  that  even  in  Richmond  some  dif- 
ferent method  may  be  evolved.  In  view  of  these  explanations,  the 
special  studies,  almost  exclusively,  have  been  directed  toward  acquiring 
information  about  and  perfecting  plans  for  mixed-refuse  destructors, 
with  the  primary  object  of  collecting  materials  at  the  lowest  expense 
and  converting  them  into  an  innocuous  product  without  causing 
nuisance  in  the  process,  and  it  seems  probable  that  the  high-tempera- 
ture system  planned  will  accomplish  the  object  desired.  Up  to  the 
present  time,  theory  and  experiment  indicate  that,  not  only  will  the 
material  collected  have  sufficient  fuel  value  in  itself  to  convert  it  into 
inoffensive  slag,  but  that,  in  addition,  there  will  be  developed  a  liberal 
amount  of  heat  units. 

In  the  installation  under  construction,  there  is  provided  a  bqiler 
which,  it  is  expected,  can  be  operated  by  the  otherwise  wasted  heat 
units,  so  as  to  furnish  ultimately  all  the  power  and  light  needed  at 
the  plant.  If  the  results  justify  the  expectations,  it  is  probable,  also, 
that  the  slag  from  the  destructor  will  eventually  be  ground  up,  and, 
with  an  admixture  of  cement,  be  converted  into  paving  blocks,  which 
would  have  value  for  gutters  and  pavement  in  places  where  traffic  is 
not  specially  heavy.  That  feature,  however,  is  only  being  considered 
for  the  future,  the  present  epoch  being  confined  to  what  might  be 
called  the  self-destruction  of  the  refuse  collected. 

If  success  attends  the  operation  of  the  new  plant,  it  is  expected 
that,  ultimately,  six  or  seven  similar  plants  will  be  installed  in  other 
portions  of  the  borough,  as  near  as  possible  to  the  centers  of  collection; 
with  one  concession,  however,  to  public  sentiment,  placing  the  de- 
structors in  manufacturing  districts,  near  a  railroad,  or  at  the  water 
front,  rather  than  in  residence  localities. 

Mr.  Fetherston's  paper  covers  the  general  phase  of  the  refuse  de- 
structor, from  the  standpoint  of  economy  in  gathering  garbage  and 
other  refuse  and  disposing  of  it  finally.  He  has  not  gone  particularly 
into  the  reasons  why  prompt  final  disposal  of  refuse  is  desirable.  There 
are,  perhaps,  throe  reasons  why  every  community  should  take  care  of 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  41 

this  feature  of  urban  life:  First,  that  which  appeals  most  popularly  Mr.  Tribus. 
to  citizens,  the  removal  of  refuse  materials  because  they  are  obnoxious 
to  the  senses  of  smell  and  sight;  second,  because  the  keeping  of  de- 
caying organic  matters  near  habitations  is  generally  supposed  to  breed 
disease;  and,  third,  a  reason  which  should  have  more  consideration, 
though  it  has  not  been  taken  up  extensively,  that,  during  the  heat  of 
the  summer,  when  the  house  fly  develops  and  feeds  and  thrives  on 
refuse,  it  is  a  very  prolific  distributor  of  disease.  It  is,  to  begin  with, 
not  a  cleanly  insect,  and  feeds  on  decaying  matter;  then,  as  likely 
as  not,  it  proceeds  to  the  nearest  receptacle  containing  milk,  for  a 
drink,  and  not  infrequently  a  bath ;  the  combination  is  often  too  much 
for  the  fly,  and  it  remains  for  a  few  moments  or  several  hours  float- 
ing around  in  the  milk,  leaving  in  it  very  often  the  germs  of  disease, 
which  in  turn  thrive  very  readily  in  the  milk  and  are  taken  into  the 
human  system.  During  the  summer,  the  human  system,  particularly 
in  infancy  and  childhood,  is  in  excellent  condition  for  the  growth  of 
disease  germs  in  the  intestines,  and  the  various  so-called  summer 
complaints  ensue.  While  probably  no  one  as  yet  will  claim  that  all 
intestinal  diseases  are  caused  by  flies;  by  the  process  of  elimination, 
in  records  that  have  been  kept  in  certain  places  by  intelligent  ob- 
servers, the  fly  can  very  fairly  be  charged  with  a  great  deal  of  the 
trouble.  The  mosquito  has  borne  its  share  of  public  contumely  as  a 
dispenser  of  yellow  fever;  why  should  not  the  ordinary  house  fly  be 
given  credit  for  the  work  which  undoubtedly  it  can  do,  and  which 
many  are  beginning  to  believe  it  does  do  ?  If  this  is  the  case,  the 
community  that  promptly  removes  and  disposes  of  its  decaying 
organic  matters  should,  first,  enjoy  the  presence  of  a  lessened  num- 
ber of  flies,  and,  second,  a  lessened  number  of  cases  of  intestinal  disease. 
This  subject  is  only  mentioned  here  as  one  worthy  of  fuller  investiga- 
tion, rather  than  as  a  conclusion  based  upon  observation. 

This  whole  topic  of  refuse  collection  and  disposal  is  one  of  very 
great  interest,  and  is  a  field  as  yet  almost  untouched  in  the  United 
States,  and,  prior  to  the  publication  of  this  paper,  very  little,  of  much 
real  value  and  based  on  facts,  has  been  printed.  The  speaker  hopes 
that  additional  information  will  be  gathered,  not  only  in  the  Borough 
of  Richmond,  but  in  other  places,  and  be  put  at  the  disposal  of  this 
Society,  to  aid  in  this  most  important  work. 

H.  Norman  Leask,  Esq. — The  speaker,  being  conversant  with  the  Mr.  Leask. 
literature  on  this  subject,  and  having  had  long  experience  in  design- 
ing and  operating  destructor  plants,  ventures  the  opinion  that,  for 
engineers,  this  is  one  of  the  most  valuable  papers  which  has  ever  been 
presented.  It  is  the  more  valuable  as  it  enters  a  new  field  and  presents 
data  from  which  a  contracting  engineer  can  design  plants  and  guaran- 
tee results  without  risk  of  failure  to  either  of  the  contracting  parties. 

It  is  pleasing  to  note  that  the   author  has  commenced  with   first 


42  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Bir.  Leask.  principles,  and  not  at  some  place  in  the  middle  of  the  subject,  which, 
unfortunately,  is  often  done.  The  exact  figures  in  the  paper,  however, 
do  not  apply  generally,  and  must  be  used  with  the  discrimination  born 
of  experience,  due  allowance  being  made  for  losses,  after  the  manner 
set  forth  in  the  problematic  balance  sheet.  Table  9. 

The  information  in  the  paper  has  not  been  available  heretofore  in 
such  extended  form;  and,  as  far  as  the  speaker  is  aware,  it  has  not 
covered  such  a  long  period,  or  such  a  great  weight  of  refuse.  At  the 
same  time,  it  should  be  remembered  that  an  inspection  of  material  is 
desirable,  in  order  to  note  any  peculiarities  in  its  character,  without 
which   immediate  success  is  not  likely  to  result. 

When  the  author  took  up  the  question,  the  conditions  existing  in 
the  Borough  of  Richmond  were  practically  the  same  as  those  which 
have  induced  most  cities  to  resort  to  destruction  by  fire.  In  many 
cities  abroad  other  methods  have  been  tried,  such  as  reduction,  making 
fertilizer,  and  gasification  for  lighting  and  power  purposes,  all  of 
which  have  failed  signally  to  deal  completely  and  in  a  satisfactory 
manner  with  the  final  disposition  of  refuse  of  all  classes,  which  is 
the  chief  desideratum.  The  only  system  of  final  disposal  which  is 
growing  in  use,  and  is  now  quite  general,  is  the  destruction  or 
cleansing  of  refuse  by  fire,  thus  rendering  it  innocuous. 

That  an  examination  of  existing  garbage  crematories  in  the  United 
States  should  offer  no  hope  of  meeting  the  requirements  satisfactorily, 
is  not  surprising,  for  such  crematories  can  hardly  be  termed  engineer- 
ing propositions,  and  one  doubts  very  much  whether  American  engi- 
neers have  had  anything  to  do  with  their  design.  The  principal  faults 
which  one  recognizes  in  garbage  crematories  of  the  present  type  are: 

1. — That  the  process  of  destructive  distillation,  rather  than 
oxidization,  has  been  resorted  to. 

2.^That  apparently  no  lower-limit  temperature  has  been  re- 
garded as  a  standard  by  the  builders  of  such  apparatus, 
although  it  is  absolutely  necessary  to  maintain  a  tem- 
perature of  more  than  1 250°  fahr.,  in  order  to  insure 
the  combustion  of  the  hydro-carbons  and  the  dissociation 
and  oxidization  of  objectionable  chemical  compounds. 

3. — That  the  usual  method  of  feeding  and  stoking  precludes  the 
possibility  of  obtaining  anything  like  a  regular  tempera- 
ture in  the  furnace,  the  temperature  rising  and  falling 
with  an  amplitude  of  probably  800°,  and  sometimes  fall- 
ing as  low  as  atmospheric  temperature. 

4. — That  the  high  temperature  is  usually  at  the  wrong  end  of 
the  furnace,  namely,  that  farthest  from  the  outlet,  and 
as  long  as  this  remains  there  can  be  no  hope  of  dealing 
successfully  with  the  material  in  a  sanitary  manner.  At- 
tempts  to   overcome   this   difficulty   have   been   made   by 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  43 

following  the  ideas  of  Mr.  Charles  Jones,  of  London,  who,  3tr.  Leask. 
in  1885,  introduced  a  fume  cremator  between  the  furnace 
and  the  chimney.     This  was  a  palliative  rather  than  a 
cure,  and'  while  it  succeeded  in  reducing  the  nuisance  to 
some  extent,  it  only  went  half  way. 

5. — Another  error,  in  certain  types  of  garbage  cremators,  is 
made  in  the  environment  of  the  burning  mass.  Water- 
jacketed  furnaces  are  absolutely  unsuitable  for  burning 
garbage  or  other  material  high  in  hydro-carbons.  In 
such  a  furnace,  flame  is  no  sooner  generated  than  it  is 
extinguished  by  absorption,  due  to  contact  with  cold  sur- 
faces, or  by  radiation.  No  one  would  think  of  hatching 
eggs  out  in  an  ice  box. 

6. — Finally,  restrictions  as  to  the  amount  of  organic  matter  re- 
maining in  the  ash  after  cremation  do  not  seem  to  be 
imposed  upon  the  builders  of  such  apparatus,  nor  have 
such  apparatus  succeeded  satisfactorily  in  eliminating 
the  organic  matter  from  the  ash. 

Table  1  is  likely  to  give  a  very  erroneous  impression  as  to  the 
quality  of  the  refuse  collected  in  Great  Britain,  That  the  author 
does  not  rely  on  the  figures  given  in  this  table,  is  quite  apparent.  His 
estimates  of  the  character  of  the  refuse  in  various  cities,  as  given  in 
Table  10,  prove  that  the  conditions  at  the  plants  he  visited  did  not 
correspond  with  the  figures  in  this  table.  Hutton's  figures  as  to  the 
percentage  of  coal,  coke,  breeze,  and  cinder  are  much  too  high, 
even  for  mid-winter.  Mr.  Codrington's  figures  appear  to  be  much 
nearer  to  the  actual  conditions,  while  those  of  Mr.  Eussell,  giv- 
ing 64.53%  for  coal,  coke,  breeze  and  cinder,  are  not  justified  by  the 
results  which  he  has  obtained  at  the  Shoreditch  plant,  the  operation  of 
which  he  directs.  The  figures  for  Torquay  appear  to  be  inverted,  and, 
if  inverted,  would  more  fairly  represent  the  conditions  existing  in 
that  town  and  in  similar  towns  along  the  south  coast  of  England  and 
in  suburban  districts. 

That  there  is  a  marked  similarity  between  the  refuse  collected  in 
Great  Britain  and  that  collected  in  New  York  is  undoubted,  and  the 
difi'erence  relates  more  to  character  than  to  calorific  value.  The 
speaker  agrees  with  the  author  that  there  is  probably  more  moisture 
in  refuse,  as  collected  in  Great  Britain,  than  in  refuse  collected  in  the 
Borough  of  Richmond,  biit  it  is  in  a  different  form.  In  the  Borough 
of  Richmond  the  moisture  is  principally  contained  in  the  garbage, 
while  in  Great  Britain  the  ash,  as  a  rule,  contains  quite  a  large  per- 
centage of  water,  and  it  must  be  remembered  that,  in  this  form,  it  is 
more  easily  attacked  than  when  carried  in  the  structure  of  material 
such  as  garbage. 


44  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Mr.  Leask.  There  is  a  similarity,  also,  between  the  refuse  collected  in  many- 
cities  on  the  Continent  of  Europe  and  that  collected  in  the  United 
States.  It  has  been  stated  that  the  refuse  collected  in  Berlin  has  a 
calorific  value  of  about  2  000  B.  t.  u.,  while  at  Frankfort  it  has  a 
calorific  value  of  4  350  B.  t.  u. ;  the  refuse  collected  in  Vienna  is 
stated  to  contain  about  3  000  B.  t.  u.,  and  that  at  Kiel  somewhat  less. 
The  refuse  collected  in  Paris  has  been  analyzed  frequently,  and  has 
been  variously  stated  to  contain  from  3  600  to  5  400  calories. 

It  should  be  noted  that  the  chemical  analysis  of  the  refuse  at  Kings 
Norton  was  made  of  refuse  collected  in  winter.  In  the  spring,  three 
years  later,  another  analysis  was  made,  and  the  refuse  was  found  to 
contain  4  300  B.  t.  u.,  In  summer,  however,  the  calorific  value  could 
not  be  much  more  than  3  000  B.  t.  v.. 

It  might  be  interesting  to  give  the  calorific  value  applied  to  various 
classes  of  refuse  by  German  scientists,  in  order  that  a  comparison  may 
be  made  between  that  part  of  Table  1  devoted  to  that  subject  and 
the  values  ascertained  by  Mr.  Welton : 

Vegetable  matter  ...  2  165  B.  t.  u. 

Bones    540        " 

Paper    3  950        " 

Sawdust   5  750        " 

Wood  6  280 

Straw    5  400        " 

Coal,  coke   9  380 

Hair    1  620 

Eags    3  600        " 

It  will  be  seen  that  the  calorific  value  of  vegetable  matter  corre- 
sponds very  closely  with  Mr.  Welton's  figures,  while  that  for  coal,  coke, 
etc.,  is  somewhat  higher,  and  that  for  rubbish  (composed  of  paper, 
wood,  rags,  etc.)  is  appreciably  lower. 

In  order  to  make  a  comparison  of  the  refuse  collected  in  the  Bor- 
ough of  Kichmond  and  that  collected  in  the  London  residential  dis- 
trict, containing  some  stores,  and  a  suburb  of  one  of  the  large 
provincial  towns,  the  speaker's  firm,  by  the  courtesy  of  the  city  en- 
gineers, made  a  number  of  analyses  of  the  refuse  as  collected  in  the 
Metropolitan  Borough  of  Stoke  Newington  and  Kings  Norton,  near 
Birmingham.    This  refuse  was  sorted  by  hand  into  four  classes: 

1. — Garbage ; 

2. — Coal,  coke,  cinders  and  fine  dust,  including  fine  inseparable 

vegetable  matter; 
3.— Eubbish; 
4. — Large  incombustible  matter,  such  as  tin,  bottles,  etc. 

These  analyses  were  made  at  a  period  corresponding  to  the  critical 
month  in  Richmond,  that  is,  September  and  early  October.    The  volume 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  45 

of  the  refuse  at  Stoke  Newington  worked  out  to  about  4  cu.  yd.  per  Mr.  Leask. 
ton  of  2  240  lb.,  and  it  contained  on  an  average :  34.43%  of  garbage, 
42.92%  of  coal,  fine  dust,  etc.,  15.4%  of  rubbish,  and  7.35%  of  glass, 
metals,  etc.  The  refuse  had  averaged  4  cu.  yd.  to  the  ton  for  about 
five  months,  and  presented  somewhat  similar  characteristics  during 
this  period.  At  Kings  Norton  the  refuse  had  a  volume  of  about  3.75 
cu.  yd.  per  ton,  and  carried  39.5%  garbage,  45.4%  coal,  fine  dust,  etc., 
9.3%  rubbish,  and  5.2%  glass,  metals,  etc.  Taking  September  alone, 
there  was  49.37%  garbage,  38.80%  coal,  coke,  etc.,  7.73%  rubbish, 
4.28%  glass,  etc.,  from  which  it  can  be  seen  that  the  percentage  of 
garbage  is  even  higher  than  that  collected  in  the  Borough  of  Richmond 
for  that  month.  It  is  not  suggested  that  the  foregoing  figures  are  ab- 
solutely accurate,  but  merely  the  result  of  an  honest  endeavor  to  ascer- 
tain the  make-up  of  the  refuse.  The  results  obtained  with  the  refuse 
at  Kings  Norton  agree  very  well  with  the  balance  sheet,  as  shovsm  in 
Table  9,  as  to  evaporation  and  combustion-chamber  temperature.  In 
September  the  evaporation  was  somewhat  higher  than  that  calculated  in 
the  balance  sheet,  that  is,  more  than  1.25,  actual,  in  ordinary  work,  and 
the  average  temperature  in  the  combustion  chamber  about  150°  higher. 
The  speaker's  make-up  of  a  balance  sheet  would  differ  slightly  in  that 
the  radiation  loss  would  not  be  as  high,  while  there  would  be  a  some- 
what lower  percentage  of  tmburned  carbon  in  the  clinker  and  ashes, 
but  probably  a  greater  loss  in  moisture  in  chimney  gases.  As  the 
average  combustion-chamber  temperature  and  evaporation  in  winter 
and  spring  are  considerably  higher  than  the  foregoing,  it  can  be  seen 
that,  in  British  refuse,  as  well  as  that  under  discussion,  there  is  a 
considerable  seasonable  variation. 

The  practical  tests,  as  given  in  Table  8,  demonstrate  clearly  that 
the  material  is  burnable,  and  the  results  obtained  are  such  as  might 
be  expected  when  burning  material  in  such  a  crude  furnace,  and  pre- 
clude all  doubt  of  obtaining  satisfactory  temperatures  with  a  prop- 
erly regulated  and  heated  air  supply. 

Based  on  the  results  obtained  with  summer  refuse  in  England, 
Messrs.  Heenan  and  Froude,  Ltd.,  of  Manchester,  who  have  been  en- 
trusted with  the  plant  for  New  York  City,  specially  designed  and 
erected  a  plant  in  Vancouver,  B.  C.  This  plant  has  certain  de- 
partures from  their  standard  type.  It  is  now  in  successful  operation, 
and,  of  the  refuse  collected  in  that  city,  it  is  destroying  more  than 
50  tons  per  24  hours,  at  suitable  temperatures,  without  the  aid  of 
supplementary  fuel,  and  with  an  excellent  residual.  The  refuse  burned 
is  very  poor  in  quality,  due  to  the  presence  of  considerable  quantities 
of  wood-ash  and  moisture.  The  difficulties  of  obtaining  high  tem- 
peratures with  this  material  have  been  overcome  by  checking  the 
quantity  and  increasing  the  temperature  of  the  air  supply.  Figures 
in  detail  as  to  the  results  are  not  yet  at  hand,  therefore  they  cannot 


46  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Mr.  Leask.  be  given  here.  The  cost  of  destruction  is  36,1  cents  per  ton,  includ- 
ing the  salaries  of  two  engineers  who  look  after  the  pump,  boiler,  fan, 
etc.  During  the  first  week  the  plant  was  put  in  operation,  it  sur- 
passed the  guaranty,  which  is  unusual — the  men  being  untrained — and 
better  results  may  be  expected  in  the  course  of  time. 

The  utilization  of  the  steam  and  clinker  resulting  from  the  de- 
struction of  refuse  is  by  no  means  the  only  offset  to  the  cost  of  burn- 
ing. The  most  important  offset  is  the  reduction  of  the  cost  of  collec- 
tion, for  a  modern  plant  may  be  placed  in  the  center  of  a  city  or 
residential  quarter  without  fear  of  nuisance  to  the  neighborhood. 
This  means  a  great  reduction  in  the  cost  of  collection  and  transport. 
Numerous  instances  can  be  cited  supporting  this:  in  the  Metropolitan 
Borough  of  Stoke  Newington,  previously  mentioned,  two  45-ton  units, 
each  with  200-h.p.  boiler  and  the  appurtenances  thereto,  a  clinker- 
crushing  and  screening  plant,  has  been  erected  in  the  middle  of  the 
Borough,  at  the  rear  of  the  Town  Hall,  and  surrounded  on  all  sides 
by  three  and  four-story  dwellings  of  a  good  class.  Notwithstanding 
the  fact  that  the  interest  and  sinking  fund  on  the  capital  outlay,  the 
repairs,  maintenance,  and  labor  charges  have  to  be  added  to  the  cost 
of  disposal,  the  cost  of  collection,  transport,  and  final  disposal  is  now 
lower  than  it  was  prior  to  the  erection  of  the  plant,  and  this  in  spite 
of  the  fact  that,  as  yet,  no  use  has  been  made  of  the  steam  generated, 
which  is  equivalent  to  about  175  k.w.  per  hour  from  one  unit;  also 
exclusive  of  the  sale  of  clinker,  which  has  been  contracted  for  on  a 
profitable  basis.  The  same  conditions  apply  to  Woodgreen,  London, 
where  the  plant  is  also  critically  located,  and  at  Rathmines,  Dublin, 
where  a  saving  in  coal  of  $2  000  per  year  (in  addition  to  the  saving 
in  cost  of  collection  and  transport)  has  been  made,  and  where,  in  the 
past,  the  power  has  been  utilized  without  relying  on  storage  batteries. 
Now  that  storage  batteries  have  been  installed,  the  saving  in  coal,  as 
shown  by  the  working  during  the  past  few  months,  will  be  more  than 
$5  000  per  anum.  This  is  the  more  remarkable  as  the  quantity  of  ref- 
use to  be  handled  is  less  than  35  tons  per  day. 

Table  9  gives  the  average  temperature  which  may  be  expected  in 
the  combustion  chamber  at  various  seasons;  it  gives  no  idea,  however, 
of  the  lowest  temperature  which  may  occur.  When  burning  Septem- 
ber refuse  on  the  standard  British  furnace  with  air  heated  to,  say, 
250°  fahr,,  the  lowest  temperature  would  be  near,  if  not  actually  be- 
low, the  limit  of  1 250°,  momentarily.  To  insure  the  temperature 
always  being  above  this  lower  figure,  it  is  absolutely  necessary  to  in- 
crease the  temperature  of  the  preheated  air,  to  control  the  air  supply 
very  carefully,  and,  further,  to  increase  temporarily  the  temperature 
of  the  air  entering  the  furnace  immediately  after  a  fresh  charge. 
Fortunately,  this  can  be  effected  by  taking  the  heat  out  of  the  clinker, 
just  withdrawn  from  the  grate,  prior  to  charging.     All  these  points 


Papers.]  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  47 

have  been  given  special  attention  in  the  case  of  the  plant  for  New  Mr.  Leask. 
York  City. 

British  destructors  have  been  designed  in  accordance  with  the 
principles  mentioned  by  the  author,  but  it  has  been  by  progressive 
steps,  and  after  many  failures.  The  various  steps  in  the  concep- 
tion and  improvement  of  refuse  furnaces,  as  made  in  England,  may 
be  traced  as  follows: 

1. — The  attempt  to  burn  refuse  in  or  under  shell  boilers; 

2. — The  building  of  a  fire-brick  lined  furnace,  or  Dutch  oven, 
operated  by  natural  draft; 

3. — The  introduction  of  the  fume  cremator; 

4. — The  abandonment  of  the  fume  cremator  and  the  introduc- 
tion of  forced  draft; 

5. — The  preheating  of  the  air  supply; 

6. — The  use  of  a  continuous  furnace  chamber,  containing  a 
number  of  grates  with  divided  ash-pits; 

7. — Suitable  ventilation  of  the  building; 

8. — Methods  of  handling  the  clinker  and  recovering  the  heat 
contained  in  it. 

With  regard  to  the  cost  of  operation,  it  is  possible,  with  a  large 
plant,  to  reduce  the  labor  charge,  part  of  the  work  being  effected  by 
mechanical  means.  It  must  be  remembered,  however,  that  the  feeding 
of  the  furnace  is  only  one  of  three  operations  necessary  in  the  work- 
ing of  the  plant :  First,  there  is  the  introduction  of  material  into 
the  furnace,  which  may  be  done  mechanically;  the  other  two,  which, 
however,  do  not  appear  to  be  susceptible  to  mechanical  operation,  are 
the  stoking  and  spreading  and  the  final  cleaning  out  of  the  mineral 
residual  from  the  grate. 

The  speaker's  firm  has  attempted  to  solve  this  problem,  and  after 
many  failures  has  at  last  succeeded  in  devising  a  machine  which  will 
handle  all  classes  of  refuse  and  will  feed  the  refuse  in  any  desired 
quantity.  Extended  trials  of  such  an  apparatus  have  been  made  at 
one  plant,  and  it  will  soon  be  installed  in  some  city.  The  system  to 
be  adopted  for  charging  depends  on  the  specific  gravity  of  the  material 
to  be  dealt  with  and  the  size  of  the  plant. 

The  speaker  has  had  the  opportunity  of  examining  the  refuse  in  a 
number  of  large  cities  in  the  United  States,  and  is  strongly  of  the 
opinion  that  the  combined  refuse  of  most  cities  can  be  destroyed  by 
fire  at  suitable  temperatures,  without  the  aid  of  supplementary  fuel. 

This  brings  up  another  phase  of  the  question  which  has  been 
mentioned  by  Mr.  Eetherston,  namely,  the  collection  of  the  refuse. 
The  adoption  of  a  system  of  a  single  collection  of  refuse,  combining 
the  ashes,  rubbish,  and  garbage,  cannot  be  urged  too  strongly.  It  is 
impossible  to  get  complete  or  satisfactory  separation.    It  is  a  question 


48  DISCUSSION  ON  MUNICIPAL  REFUSE  DISPOSAL  [Papers. 

Mr.  Leask.  of  public  health,  rather  than  profit.  The  single  collection  costs  less 
tc  make.  The  mixing  of  thd  refuse  retards  decomposition,  the  ashes 
acting  as  a  deodorant  by  absorption,  and  it  provides  a  wherewithal  in 
calories  to  cleanse  the  mass  of  its  impurities,  and,  when  burned,  leaves 
a  marketable  residual  in  the  form  of  clinker.  If  advantage  be  taken 
of  the  heat  generated  by  combustion  (and  here  it  should  be  noted 
that,  whatever  the  material  may  be,  when  it  is  burned  at  suitable 
temperatures  there  is  always  utilizable  heat),  there  is  placed  at  the 
disposal  of  the  authorities  another  valuable  residual  in  the  form  of 
steam,  the  best  uses  for  which  are  those  giving  a  large  load  factor, 
such  as  pumping — sewage  or  water — or  electric  traction.  Lighting 
alone  is  not  a  satisfactory  outlet.  The  question  of  collection  and  dis- 
posal of  refuse  in  the  United  States  to-day  appears  to  be  in  the  state 
that  it  was  in  older  countries  some  years  ago,  that  is,  in  the  hands  of 
contractors.  It  has  been  abundantly  demonstrated  that  the  only  satis- 
factory method  is  for  the  municipalities  themselves  to  undertake  it. 
Where  this  has  been  done,  and  where  refuse  destructors  form  a  part 
of  the  scheme,  it  has  been  followed  by  a  noticeable  decrease  in  the 
death  rate.  Why  should  not  one  of  the  richest  coimtries  in  the  world 
forsake  the  problematical  gain  arising  from  reduction,  and  regard  the 
question  from  a  purely  public  health  standpoint,  as  has  been  done  in 
older  countries,  even  in  backward  Russia? 

The  whole  question  is  purely  one  of  combustion,  and,  generally 
speaking — provided  the  moisture  contained  in  the  material  is  not  ex- 
cessive— refuse  containing  2  000  B.  t.  u.  per  lb.  will  cleanse  itself. 

The  great  difference  between  the  combustion  problem  as  applied 
to  coal  and  refuse  is  this :  When  dealing  with  coal  one  has  a  material 
which  is  comparatively  low  in  ash  and  requires  about  20  to  24  lb.  of 
air  per  lb.  to  burn  it  in  a  practical  and  satisfactory  manner,  whereas 
refuse  is  high  in  ash,  and  the  air  required  is  only  from  4  to  5  lb.  per 
lb.  The  difficulty,  therefore,  is  to  find  a  small  quantity  of  carbon  in 
so  large  a  bulk,  with  the  minimum  quantity  of  air.  To  effect  this 
one  must  look  to  the  distribution  and  the  temperature  of  the  air  sup- 
plied, the  intensity  of  the  draft,  and  the  environment  of  the  burning 
mass.    It  is,  therefore,  wholly  an  engineering  proposition. 

The  speaker  must  again  congratulate  the  author  on  the  service  he 
has  rendered  to  engineering  science  in  general  and  contracting  engi- 
neers in  particular. 


Vol.  XXXIV.  JANUARY,   1908.  No.  1. 


AMEEICAN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED    185  2. 


PAPERS  AND   DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


THE  EEINFORCED  CONCRETE  WORK  OF  THE 
McGRAW  BUILDING. 

Discussion.* 


By  Messrs.  Guy  B.  Waite  and  E.  P.  Goodrich. 


Guy  B.  Waite,  M.  Am,  Soc.  C.  E. — Had  the  McGraw  Building  Mr.  Waite. 
been  designed  ten  years  ago,  it  would  not  have  been  built  of  reinforced 
concrete.  Few  engineers  at  that  time  had  sufficient  confidence  to 
undertake  the  experiment,  and  the  local  building  laws  did  not  permit 
it.  Even  more  recently  than  this,  nearly  every  important  architect 
in  the  vicinity  would  have  refused  to  listen  to  an  argument  for  using 
concrete  in  a  building  in  which  there  was  to  be  heavy  vibration,  as  it 
was  only  considered  fit  for  light,  cheap  buildings. 

At  that  time,  even  engineers  were  afraid  to  speak  in  favor  of  con- 
crete for  general  building  purposes,  for  fear  of  becoming  unpopular. 
One  who  dared  engage  in  concrete  building  construction  had  to  en- 
dure the  humiliation  of  seeing  some  of  his  former  friends  pass  quietly 
to  the  opposite  side  of  the  street  when  they  saw  him,  in  order  to  avoid 
meeting  one  who  was  engaged  in  an  immoral  business,  and  who  was 
just  escaping  the  meshes  of  the  law.  The  change  in  public  opinion, 
with  regard  to  reinforced  concrete,  has  been  brought  about  purely  by 
the  merits  of  the  construction. 

About  ten  years  ago  the  Building  Department  of  New  York  City 
inaugurated  standard  tests  for  concrete  constructions  to  be  used  for 
fire-proof  floors.  All  constructions  for  floors  had  to  be  submitted  to  a 
4-hour  fire  test,  at  an  average  temperature  of  1 700°  fahr.  They 
were  to  be  loaded  with  150  lb.  per  sq.  ft.  on  a  full-sized  floor  not  less 

*  Continued  from  December,  1907,  Proceedings. 


50  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Waite.  than  14  ft.  long.  Immediately  following  the  fire — the  material  being 
still  red-hot — a  regulation  stream  of  cold  water  was  to  be  thrown  on 
the  construction,  with  a  pressure  of  60  lb.  per  sq.  in.,  for  10  min. 
The  construction  was  then  to  withstand  a  distributed  load  of  600  lb. 
per  sq.  ft." 

During  the  five  years  following  the  inauguration  of  this  test,  about 
twenty-five  concrete  constructions  withstood  it  successfully.  By  this 
time  some  of  the  public  had  been  convinced  that  concrete  possessed 
merits,  as  a  fire-proof  material,  but  did  not  dare  to  speak  out;  while 
others  feared  that  it  had  some  merits,  and  set  out  to  kill  it.  Quite 
successful  obstacles  were  placed  in  its  way,  by  the  Board  of  Insur- 
ance Underwriters,  who  fined  it;  by  codes  of  law,  which  practically 
ruled  it  out;  by  labor  unions,  who  dictated  by  whom  and  how  it  should 
be  made;  and  by  politicians,  whose  interests  were  generally  in  other 
directions. 

Interest  was  finally  awakened  by  the  favorable  showing  made  by 
concrete  in  some  of  the  great  fires  on  which  a  few  honest  reports  were 
made  by  eminent  engineers,  and  since  that  time  concrete  constructions 
have  gained  rapidly  in  popularity. 

It  was  only  about  four  years  ago  that  the  speaker  secured,  from 
the  Department  of  Buildings  of  New  York  City,  the  first  permit  ever 
granted  in  the  Borough  of  Manhattan  for  a  concrete  building,  in- 
cluding concrete  wall  construction.  All  who  are  interested  in  the 
advancement  of  reinforced  concrete  must  feel  indebted  to  Professor 
Burr  for  giving  his  name  and  influence  to  this  cause. 

The  McGraw  Building  was  undoubtedly  made  in  reinforced  con- 
crete because  it  ofi^ered  the  best  construction  to  withstand  the 
peculiarly  heavy  work  of  a  printing  house;  and  because  it  gave  the 
safest  fire  risk.  The  National  Board  of  Fire  Underwriters,  some  two 
years  ago,  recommended  a  minimum  rate  of  insurance  on  similar  con- 
structions, and,  from  recent  inquiry  of  the  Local  Board  of  Fire  Un- 
dervtrriters,  it  is  learned  that  even  this  august  body  has  at  last  reached 
a  point  where  a  similar  action  is  under  consideration. 

As  a  fire  risk,  concrete  structures  offer  the  best  possible  invest- 
ment, either  for  individuals  to  carry  their  own  risks,  or  for  insurance 
associations  to  make  a  specialty  of  these  constructions.  Such  under- 
takings would  be  the  safest  and  most  profitable  kind  of  insurance 
ventures. 

Concrete  is  superior  to  burned  clay,  not  because  it  is  more  fire- 
proof, but  on  account  of  its  superior  elasticity  under  the  stress  due  to 
fire  and  water.  Some  years  ago  the  speaker  constructed,  entirely  of 
cinder  concrete,  a  test  house,  14  by  14  ft.  and  about  12  ft.  high,  with 
walls  12  in.  thick,  and  with  a  ceiling  only  IJ  in.  thick.  In  various 
tests  conducted  by  the  Department  of  Buildings  in  that  house  the 
ceiling  was  submitted  to  four  separate  series  of  fire  tests. 


PLATE   II. 

PAPERS,  AM.  SOC.  C.  E. 

JANUARY,  1908. 

WAITE  ON 

REINFORCED   CONCRETE    BUILDINGS. 


Fig.  1.— Cinder  Concrete  Test  House,  After  Fourth  Fire 


Fig  3.— Stone  Concrete  House,  with  Test  Load, 
After  Four-Hour  Fire  Test. 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING  51 

When  this  structure  was  torn  down,  to  make  way  for  dock  im-  Mr.  Waite. 
provements,  this  IJ-in.  ceiling  was  examined  by  an  engineer  from  the 
Department  of  Buildings  of  New  York  City,  and  by  Professor  Wool- 
son  of  Columbia  University,  and  was  found  to  be  in  good  condition, 
the  fire  having  affected  scarcely  ^  in.  on  its  under  side,  and  this  was 
due  to  the  first  fire. 

While  reinforced  concrete  has  been  demonstrated  to  be  superior 
in  many  respects  to  other  forms  of  fire-proof  construction,  it  has 
forced  its  way  to  the  front  principally  on  account  of  the  economy  it 
has  effected. 

The  parts  of  a  building  in  which  it  is  best  adapted  must  be  deter- 
mined largely  by  the  engineer's  experience.  Sometimes  this  experi- 
ence is  paid  for  very  dearly,  and  forms  a  secret  chapter  in  his 
biography. 

Local  conditions  often  alter  circumstances  to  such  an  extent  that 
a  kind  of  construction  which  might  be  erected  in  one  locality  at  a 
profit    would  become  a  loss  in  another  locality  a  short  distance  away. 

As  the  relative  prices  of  built  steelwork  and  concrete  per  unit 
section  are  about  as  65  to  1,  and  the  relative  working  capacities  in 
compression  (16  000  to  500)  are  about  as  30  to  1,  it  is  evident  that, 
other  things  being  the  same,  the  more  concrete  is  substituted  for  steel 
in  compression,  the  greater  is  the  economy. 

Where  heavy  loads  are  to  be  carried,  concrete  will  be  found  highly 
advantageous;  conversely,  where  small  loads  are  carried,  it  will  have 
little  advantage.  Keeping  this  fact  in  view,  one  would  expect  to  find 
the  greatest  economy  in  using  concrete  for  column  supports  and  floor 
constructions — the  heavier  the  construction,  the  greater  the  economy. 

The  great  barrier  to  using  concrete  for  columns  is  the  impractical 
size  necessary  when  more  than  a  few  stories  are  required.  Leaving  to 
others  the  discussion  of  the  rationality  of  the  combination  of  con- 
crete and  steel  in  the  columns  of  the  McGraw  Building,  the  speaker 
considers  this  form  of  construction  siiperior  to  anything  heretofore 
done  in  reinforced  columns. 

The  column  is  reduced  to  a  reasonable  size,  and  is  made  safe 
against  accidents.  The  positive  dead  loads  from  the  building  are 
carried  by  positive  steel  supports,  while  the  doubtful  superimposed 
floor  loads  are  adequately  provided  for  by  the  more  questionable  form 
of  concrete  reinforcement. 

The  certainties  are  balanced  one  against  the  other,  and  the  un- 
certainties are  also  brought  to  face  each  other.  Although  columns  of 
this  form  are  not  directly  the  cheapest,  the  speaker  believes  them  to  be 
the  most  economical,  all  things  considered.  Columns  of  this  form 
are  adapted  to  much  speedier  erection  than  the  cheaper  reinforced 
concrete,  and  are  absolutely  safe  during  erection;  they  also  allow 
as  good  a  monolithic  connection  of  the  column  with  the  floor  as  that 


52  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  waite.  obtained  in  other  reinforced  forms.  The  column  forms,  with  the  steel 
as  a  gxiide,  cost  less  than  where  they  are  made  as  independent  struc- 
tures. 

As  pointed  out  by  the  author,  the  construction  of  the  forms  is 
probably  the  greatest  problem  in  the  practical  construction  of  rein- 
forced concrete.  Almost  every  beginner  in  this  field  has  arrogant 
confidence  in  his  ability  to  eclipse  everything  previously  done  in  the 
way  of  perfect  centering.  It  is  only  necessary  to  watch  such  an  one 
and  see  a  second  scheme  in  his  second  job,  a  third  scheme  in  his  third 
job,  and  so  on,  imtil  he  becomes  a  meek  plodder  along  the  tow-path 
of  experience. 

A  discrimination  should  be  made  between  centers  or  forms  de- 
signed for  a  building,  and  a  building  designed  for  the  forms.  Forms 
made  to  fit  a  special  building  may  cost  several  times  as  much  as  those 
with  which  a  standard  building  might  be  made.  The  cost  of  center- 
ings may  be  reduced  about  in  proportion  to  the  standardization  of 
the  building.  Many  useful  schemes  for  systematizing  the  general 
construction  of  centerings  have  been  invented,  and  many  of  these 
have  simplified  the  problem  so  that  the  main  cost  is  in  taking  down 
and  putting  up  the  forms. 

Most  of  these  centering  schemes  are  used  for  rough  concrete  work, 
where  the  surface  is  to  be  plastered  afterward;  but  when  finished  sur- 
faces are  to  be  produced,  the  cost  of  centers  is  more  than  doubled.  If 
the  mechanic  trained  to  do  finished-center  work  be  told  to  make  rough 
standard  centers,  he  will  spend  nearly  7i  hours  carefully  getting 
ready  to  do  work  which  would  take  the  other  man  J  hour;  conversely, 
the  rapid  standard-center  man  if  put  on  perfect-center  work,  would 
do  in  i  hour  what  would  take  the  perfect-center  man  about  7i  hours 
to  undo  and  do  over  again.  False  conceptions  and  misrepresentations 
on  the  part  of  competitors  in  concrete  work  have  led  them  into  bitter 
warfare  by  presenting  owners  with  what  they  term  "finished  surface" 
free  of  cost.  Is  it  not  possible  to  conceive  a  method  of  trying  uni- 
formly rough  surfaces  instead  of  uniformly  smooth  surfaces,  as  a 
help  toward  solving  the  problem  of  centers  ? 

The  indirect  method  may  be  used  to  cheapen  centers,  that  is,  con- 
structions may  be  used  which  do  not  require  the  expensive  centering 
necessary  in  ordinary  forms  of  reinforced  concrete.  If  such  construc- 
tions de  not  advance  the  total  cost  by  increases  in  other  directions, 
there  will  be  a  net  saving.  As  is  evident,  the  forms  or  centers  for 
ordinary  reinforced  concrete  must  be  sufiiciently  heavy  to  maintain 
a  perfectly ,  independent  structure  under  the  tendency  to  warp  and 
deflect,  due  to  the  fact  that  they  are  alternately  wet  and  dry,  and  also 
on  account  of  the  heavy  load  of  the  concrete.  Where  a  steel  skeleton  (such 
as  the  columns  in  the  McGraw  Building),  maintains  the  construc- 
tion lines,  one  may  use  for  centers  material  which  is  much  lighter,  and 


PLATE  III. 

PAPERS,  AM.  SOC.  C.  E. 

JANUARY,   1908 

WAITE  ON 

REINFORCED    CONCRETE    BUILDINGS. 


I^^^^IP*^^ 

i^^m 

■ 

w 

9 

^S^k 

^^^^^^Kl 

p 

I 

] 

K 

^.  ^    .. 

p 

H^^^i 

wK^^^^^ 

^^ 

Bv^^> 

E- 

"■^^^^ 

^^^^■hL 

mKi 

*'i 

'"""^■^-^^i 

Fig.  1. — The  Bonwit-Teller  Building 


Fig.  3.— The  Salvation  Army  Warehouse. 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING  53 

more  easily  worked  and  handled   than  with  ordinary  reinforced  con- Mr.  Waite. 
Crete. 

The  author  does  not  describe  the  wall  construction  of  this  build- 
ing, but  the  speaker  believes  it  to  be  of  some  form  of  reinforced  con- 
crete. In  wall  construction,  the  conditions  are  very  different  from 
either  column  or  floor  construction.  In  both  columns  and  floors,  con- 
crete makes  a  saving  in  steel,  but,  in  wall  construction,  this  element 
of  saving  does  not  enter.  Further,  in  monolithic  wall  construction 
two  forms  must  be  kept  plumb,  but  in  floors  only  one  form  is  re- 
quired, and  gravity  helps  to  hold  this  in  place.  The  speaker,  with  the 
very  best  of  assistants,  after  finishing  several  buildings  having  com- 
plete concrete  wall  construction  averaging  8  in.  thick,  concluded  that 
the  cost  of  the  concrete  in  such  wall  construction  was  of  small  con- 
sequence and  could  be  safely  neglected  in  totaling  up  the  entire  cost 
of  the  wall.  In  factory  construction,  where  there  are  practically  no 
walls  except  panels  under  windows,  no  such  difficulties  are  en- 
countered as  in  dead  wall  construction. 

Nothing  is  stated  definitely  by  the  author  concerning  the  character 
of  the  steel  used  in  the  floor  construction,  other  than  that  round  rods 
were  used.  The  kind  of  bars  and  the  character  of  the  steel  in  them 
seem  to  command  a  great  deal  of  attention  at  present.  The  only 
logical  conclusions  that  can  be  drawn  from  the  claims  of  the  big 
grist  of  deformers  (with  increased  capacities  for  each  new  deforma- 
tion in  their  rods)  is  that  they  are  developing  the  art  toward  a  state 
where  (according  to  claims)  practically  nothing  but  bond  and  grip 
will  be  required,  and  steel  for  tension,  etc.,  as  now  designed,  will  be- 
come of  little  consequence. 

In  referring  to  some  recent  constructions  which  the  speaker  has 
^xecuted,  the  only  excuse  he  has  to  offer  for  so  doing  is  that  such 
improvements  have  come  after  quite  a  lengthy  experience  in  general 
steel  and  reinforced  concrete  construction,  and,  being  a  product  of 
natural  evolution,  they  belong  to  the  general  scheme  of  development 
toward  something  higher.  The  speaker's  experiences  have  been  unlike 
those  of  many  engaged  in  reinforced  concrete  construction,  because, 
m  most  cases,  he  has  had  to  contend  with  the  conditions  existing  in 
crowded^  parts  of  large  cities,  where  space  for  storing  materials  and 
performing  work  is  extremely  limited,  and  where  great  rapidity  of 
erection  is  necessary;  and,  on  account  of  the  extra  height  of  build- 
ings, safe  construction  must  be  considered. 

Having  been  fundamentally  trained  in  steel  construction,  followed 
by  the  fire-proofing  of  the  steel;  and  subsequently  having  pursued 
general  reinforced  concrete  construction,  the  speaker  was  forced  to 
consider  the  merits  and  demerits  of  the  combinations  of  these  three 
factors  in  building  construction  for  the  conditions  found  in  large 
cities. 


54  DISCUSSION  ON  EEINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Waite.  The  Safety  and  the  speed  of  steel  construction  were  apparent,  and 
the  advantages  in  the  use  of  concrete  for  the  protection  and  fire- 
proofing  of  steel  were  well  demonstrated.  Then  followed  the  combina- 
tion of  steel  with  concrete  (formerly  used  for  fire-proofing),  and  this 
developed  into  a  system  which  possessed  all  the  merits  of  the  steel 
skeleton  construction  and  the  advantages  of  reinforced  concrete.  In 
this  system  (known  as  System  "M,"  in  its  order  with  other  systems) 
there  is  required  only  from  35  to  40%  of  the  steel  necessary  for  the  con- 
ditions in  which  the  steel  does  all  the  work.  The  concrete — which 
must  be  used  for  fire-proofing — is  made  to  do  the  remainder.  The 
light  steel  frame  is  run  up  ahead  of  the  concrete,  in  the  usual  manner 
for  steel  frames,  and  is  made  strong  enough  to  take  all  tensional  and 
shearing  stresses  in  the  subsequently  reinforced  construction  formed 
by  the  steel  and  concrete.  The  combination  forms  a  truly  reinforced 
structure. 

Work  can  be  done  on  several  stories  simultaneously,  as  in  other 
steel  construction.  The  necessary  forms  are  simplified,  as  compared 
with  those  required  in  most  reinforced  concrete  constructions,  on  ac- 
count of  the  assistance  given  by  the  steel  frame.  Within  the  last  two 
years,  some  twenty  buildings  in  the  vicinity  of  New  York  City  have 
been  constructed  by  this  system. 

From  January  1st  to  April,  1907,  while  the  McGraw  Building  was 
being  erected,  three  buildings  in  that  vicinity  were  constructed  in 
which  the  floors  were  of  this  form  of  construction,  namely:  The 
Bonwit  Teller  Building,*  at  15  and  17  West  Thirty-fourth  Street, 
shown  by  Fig.  1,  Plate  III;  the  Salvation  Army  Warehouse,!  at  533 
and  537  West  Forty-eighth  Street,  shown  by  Fig.  2,  Plate  III ;  and  the 
Strack  Building,  at  214  and  220  West  Twenty-third  Street. 

In  these  buildings  steel  columns  carry  the  entire  loads,  but,  where 
conditions  permit,  a  light  steel  frame,  similar  to  the  construction  used 
in  the  McGraw  Building,  carries  the  dead  floor  loads. 

The  general  form  of  the  steel  in  this  combined  column  construc- 
tion is  shown  by  Fig.  1.  It  is  made  of  channels  or  similar  sections 
disposed  centrally  with  respect  to  floor  beams  and  girders,  and  the 
separate  steel  members  are  connected  at  the  corners.  The  speaker 
has  found  this  to  give  a  very  simple  and  effective  steel  skeleton  which, 
he  believes,  affords  ample  means  for  the  proper  combination  of  the 
steel  and  concrete. 

Some  months  ago,  in  building  a  garage  for  the  use  of  his  family, 
at  Whitestone,  Long  Island,  the  speaker  concluded  to  make  the  entire 
building  of  reinforced  concrete  in  order  to  demonstrate  the  economy 
of  a  new  form  of  construction  suitable  for  a  small  number  of  laborers. 
The  building  is  40  ft.  long  and  20  ft.  wide,  and  has  two  stories,  and 

*  Engineering  Neirs,  April  25th,  1907. 

+  The  Engineering  Record,  June  '23d,  1907. 


PLATE  IV. 

PAPERS,  AM.  SOC.  C.  E. 

JANUARY,  1908. 

WAITE  ON 

REINFORCED   CONCRETE    BUILDINGS. 


Reinforced  Concrete  Garage  at  VVhitestone,  Long  Island 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING 


55 


an  attic.  The  walls  consist  of  a  series  of  12  by  8-in.  reinforced  Uv.  waite. 
pilasters,  spaced  5  ft.  apart  between  centers.  Between  these  buttresses, 
and  erected  simultaneously  with  them,  there  are  concrete  blocks,  3  in. 
thick  and  12  in.  wide.  In  the  lower  story  the  blocks  are  flush  with 
the  outside  of  the  pilasters,  and  in  the  second  story  they  are  kept  back 
from  the  front  to  give  the  effect  shown  in  the  photograph,  Plate  IV. 


m 


jSi 


W^ 


I 


-^. 


SYSTEM  "M"  COLUMN 
Fig.  1. 

The  floors  were  reinforced  with  4  by  T^-in.  steel  beams,  resting  on  the 
pilasters,  and  having  shear  bars  extending  up  from  holes  in  the  webs 
of  the  beams.  One  handy  man  and  two  laborers  constructed  the 
foundations  and  all  the  walls  and  floors  in  about  8  weeks.  The  walls 
were  run  up  several  feet  above  the  second  floor  in  order  to  make  a 
full  story  of  the  attic. 

E.  P.  Goodrich,  M.  Am.  Soc.  C.  E. — The  speaker's  connection  with  Mr.  Goodrich, 
the  McGraw  Building,  in  a  supervisory  capacity,  during  the  major 
portion  of  its  design  and  construction  makes  Professor  Burr's  descrip- 
tion of  special  interest  to  him.  In  a  few  particulars,  that  description 
may  be  somewhat  amplified,  for  the  sake  of  noting  additional  points 
of  interest. 

The  power  plant  for  the  building  is  located  in  a  sub-basement 
situated  in  the  southwest  corner.  Consequently,  the  columns  in  that 
portion  of  the  building  are  somewhat  langer  than  the  others  and  ex- 
ceed the  dimensions  given  in  the  paper  by  12i  ft.,  making  the  length     . 


56  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Goodrich,  of  the  longest  column  172  ft.  It  may  also  be  of  interest  to  note  that 
the  "reinforcement"  in  the  first  length  of  one  of  these  columns  weighed 
14  050  lb. 

The  windows  on  the  sides  and  rear  of  the  building  are  of  wire-glass 
in  metal  frames,  so  that  practically  the  only  possible  additional  device 
which  could  be  added  to  give  security  against  fire,  would  be  a  com- 
plete automatic  sprinkler  system.  In  consequence,  the  McGraw  Build- 
ing is  one  of  the  best  in  the  city,  as  far  as  insurance  conditions  are 
involved,  and  carries  a  very  low  rate  for  both  the  building  and  the 
contents. 

The  column  spacing  was  determined  primarily  by  the  dimensions 
of  modern  printing  presses,  nearly  half  a  score  of  which  are  now  in 
operation  on  several  of  the  upper  floors  of  the  building.  This  fact 
brought  about  the  use  of  rectangular  floor  bays,  while  a  more  nearly 
square  arrangement  would  have  been  slightly  more  economical,  had  it 
been  possible  to  design  the  building  in  that  way. 

Of  the  several  new  features  in  the  building,  of  course,  the  column 
design  is  the  most  unusual.  While  the  whole  arrangement,  as  finally 
worked  out,  proved  highly  satisfactory,  from  a  construction  point  of 
view,  it  may  be  open  to  some  adverse  criticism,  from  a  solely  economic 
standpoint.  As  shown  by  Mr.  Douglas,  a  design  for  purely  structural 
columns  would  have  cost  less  money,  and  Mr.  Stern  suggests  that,  even 
when  fire-proofed,  such  columns  wovild  have  been  smaller  than  those 
used.  Plenty  of  evidence  has  been  adduced  from  the  San  Francisco 
conflagration  to  show  that  no  comparison  can  be  made  between  struc- 
tural columns,  however  well  "fire-proofed"  in  the  usual  manner,  and 
such  columns  as  are  used  in  the  McGraw  Building.  No  comparison 
is  fair  unless  this  superiority  to  resist  fire  is  capitalized.  On  the 
other  hand,  columns  of  the  Considere  type  likewise  possess  this  good 
quality,  their  principal  drawback,  under  such  conditions,  being  their 
size.  It  may  be  of  interest  to  state  that,  early  in  the  history  of  the 
design  of  this  building,  the  speaker  caused  to  be  prepared  a  typical 
column  of  the  Considere  type,  based  on  the  accepted  stress  require- 
ments of  the  New  York  City,  Manhattan  Borough,  Building  Regula- 
tions at  that  time.  The  columns,  of  course,  were  circular  in  section, 
with  a  diameter  in  no  case  greater  than  the  diagonal  of  the  corre- 
sponding square  column  of  the  Burr  type,  finally  vised.  The  estimated 
cost  of  the  Considere  column,  on  a  conservative  basis,  showed  an  ap- 
parent saving  in  its  favor  of  approximately  $10  000  for  the  whole 
building. 

Two  other  small  objections  to  the  Burr  column  were  also  discovered, 
which  wore  almost  entirely  obviated  during  the  progress  of  the  work, 
and  could  be  entirely  remedied  in  future  designs.  The  wide  faces  of 
the  angles  in  the  lower  stories,  and  the  wider  expanses  of  some  of  the 
splice-plates,  made  necessary  a  special  wrapping  of  wire  or  wire  lath 


Papers.]        DISCUSSION  ON"  REINFORCED  CONCRETE  BUILDING  57 

to  hold  the  fire-proofing  concrete  in  place ;  and  the  extreme  rigidity  Mr.  Goodrich, 
of  the  column  steel,  made  necessary  a  much  more  careful  adjustment 
of  the  forms  than  is  usually  required  for  reinforced  concrete  build- 
ings. In  most  cases,  the  less  rigid  reinforcing  rods  are  given  slight 
eccentricities,  which  do  not  affect  their  efficiency  seriously,  and  are 
thus  made  to  accommodate  themselves  to  small  variations  in  the  spac- 
ing of  the  forms,  and  thereby  save  some  labor  cost.  This  latter  possi- 
ble defect  of  the  rigid  reinforcement  may  even  be  considered  a '  real 
virtue,  in  the  eyes  of  some  people. 

The  speaker  is  aware  of  really  very  few  reliable  tests  of  reinforced 
concrete  columns,  and  of  none  which  possessed  anything  like  the  per- 
centage of  longitudinal  steel  found  in  those  of  the  McGraw  Building. 
Some  time  ago,  the  speaker  arranged  for  a  series  of  tests  on  specimens 
designed  after  the  Burr  type,  and  of  practically  full  size,  but,  un- 
fortunately, the  results  have  not  yet  been  secured.  In  this,  it  is  well 
to  note  a  fact  to  which  Professor  Morsch,  of  Zurich,  calls  attention, 
in  his  "Eisenbetonbau,"  that  the  efficiency  of  longitudinal  rod  rein- 
forcement decreases  with  the  increase  of  the  percentage  used,  at  least 
up  to  4%,  and  that  no  one  knows  how  larger  amounts  will  act.  It  is 
thus  incumbent  upon  designers  to  exercise  great  care  in  selecting 
working  stresses  for  concrete  columns  possessing  considerable  longi- 
tudinal steel,  as  the  field  is  absolutely  unknown  at  the  present  time, 
and  some  serious  trouble  may  result  for  inexperienced  designers  who 
follow  rules  blindly. 

Another  point  to  be  noted  is  the  fact  that  most  experimenters  on 
concrete  columns  have  concluded  that  the  concrete  appears  to  carry 
much  the  larger  percentage  of  the  load  until  it  has  reached  a  stress 
far  above  the  usual  allowable  working  one,  when  the  steel  comes  into 
more  pronounced  action.  Of  course,  this  conclusfon  is  based  on  com- 
putations involving  an  assumed  modulus  of  elasticity  of  the  steel  and 
the  observed  stresses  and  strains  of  the  column.  The  distribution  of 
stress,  above  described,  is  probably  due  to  the  fact  that  the  stress- 
strain  curve  for  concrete  is  not  a  straight  line,  thus  demonstrating  the 
existence  of  a  variable  modulus  of  elasticity.  From  these  facts,  it 
might  seem  to  be  more  rational  to  reverse  the  condition  as  to  the  dead 
and  live  load  carrying  capacities  of  the  steel  and  concrete  in  the  Burr 
column,  and  require  the  concrete,  at  say  750  lb.  unit  stress,  to  carry 
all  the  dead  load  and  then  add  enough  steel  in  structural  form,  if  so 
desired,  to  carry  the  total  or  reduced  live  loads. 

Another  item  of  design  in  the  McGraw  Building  to  which  special 
attention  was  paid,  was  the  connection  between  the  reinforcing  rods 
and  the  column  steel.  This  was  worked  out  so  effectively  that  the 
steel  erectors  of  the  columns  often  attached,  to  any  convenient  point 
of  the  beam  reinforcement,  one  end  of  the  turnbuckle  which  they  used 
for  plumbing  the  column  sections.  In  no  other  reinforced  concrete 
building  within  the  speaker's  knowledge   could  this  be  done. 


58  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Goodrich.  With  such  rigidity  of  column  steel  and  its  firm  connection  with 
the  beam  rods,  the  best  method  of  beam  design  would  seem  to  be  that 
of  cantilevers  or  continuous  beams  throughout,  instead  of  simply  sup- 
ported members.  Such  an  arrangement  of  the  steel  in  a  concrete 
beam  has  the  following  advantages: 

Maximum  shears  occur  at  points  where  maximum  moments  are 
found,  and,  in  consequence,  where  most  steel  is  placed. 

Not  as  much  steel  is  found  near  the  bottoms  of  beams,  where  it 
would  be  exposed  to  the  most  trying  effects  of  fire. 

Such  a  method  of  design  obviates  the  tendency  to  sharp  deflec- 
tions near  the  supports,  with  the  resulting  probability  of 
the  occurrence  of  cracks  at  points  where  the  shear  is  the 
greatest. 

Such  design  gives  most  resistance  against  the  type  of  failure 
observed  in  impact  experiments. 

There  is  also  less  likelihood  of  the  displacement  of  reinforcement, 
because  it  is  in  view  during  the  greater  part  of  the  process 
of  concreting. 

All  beams  and  girders  throughout  the  McGraw  Building  were  de- 
signed as  fully  continuous,  or  restrained,  even  where  supported  in  the 
outside  columns  and  walls.  The  drawings  show  twice  as  much  steel 
over  tlie  supports  as  in  the  centers  of  the  spans,  and,  since  the  factor 
used  in  connection  with  the  moment  at  the  latter  point  is  xjj,  according 
to  the  requirements  of  the  Building  Code,  the  factor  for  the  supports 
is  only  i.  Thus  it  is  seen  that,  when  compared  with  ^^^,  and  ■^^. 
the  theoretically  correct  vakies,  more  than  twice  as  much  beam  and 
girder  reinforcement  was  used  as  theory  would  dictate.  This  extra 
material  was  used  iji  a  literal  compliance  with  the  anomalous  wording 
of  the  New  York  Building  Code.  A  comparison  of  this  building  with 
numerous  others  has  led  the  speaker  to  the  conclusion  that  the  re- 
quirements therein  contained  are  rarely  complied  with  literally,  and 
that  this  faulty  requirement  of  the  code  has  been  the  real  cause  of 
much  poor  work. 

To  the  speaker,  no  reason  is  apparent  for  using,  over  points  of 
support,  more  steel  than  enough  to  satisfy  the  theoretical  moment 
formula,  with  a  coefficient  of  y\.  When  that  amount  is  used  in  that 
way,  only  half  as  much,  of  course,  is  theoretically  necessary  in  the 
lower  part  of  a  beam  at  its  center,  while  the  building  requirements 
specify  as  much  as  would  be  indicated  by  a  coefficient  of  r^,  which  is 
even  more  than  is  needed  in  the  upper  part  of  the  beam  over  a  support. 

Some  slight  argument  may  be  advanced  for  using  as  much  steel 
below  as  above,  in  the  two  locations,  from  the  fact  that  eccentrically 
placed  partial  loads  on  continuous  members  resting  on  perfectly  mov- 
able supports,  subject  the  members  to  maximum  positive  and  negative 
moments  which  are  much  larger  than  those  produced  by  a  continuous 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING  59 

load,  as  usually  considered.  This  fact  has  often  led  the  speaker  to  Mr.  Goodrich, 
recommend  a  partial  concession  to  the  older  ideas  of  design,  and  to 
use  equal  amounts  of  steel  over  the  supports  and  at  the  centers  of 
spans,  determining  this  quantity  by  the  coefficient,  tV-  This  distri- 
bution allows  of  an  economical  design  for,  and  method  of  handling, 
the  rods;  it  meets  practically  all  the  requirements  of  partial  loads, 
and,  at  the  center  of  the  spans,  uses  within  20%  of  the  quantity  of 
steel  required  by  the  New  York  Building  Code,  with  100%  better  dis- 
tribution, as  far  as  prevention  of  cracks  is  concerned. 

The  reinforced  concrete  beams  and  girders  of  a  monolithic  con- 
crete building  are  not  beams  and  girders  at  all,  in  the  sense  of  the 
wooden  and  steel  ones  in  the  older  types  of  structures,  which  simply 
rest  on  brackets  and  have  ample  opportunity  for  motion  in  each  joint. 
Until  cracks  have  formed,  the  concrete  beams  are  really  extended 
brackets  on  the  columns  and  other  members,  and  should  be  designed 
as  such.  The  early  workers  with  reinforced  concrete  were  influenced 
too  largely  by  the  old  type  of  structure,  and  few  designers  have  even 
yet  grown  into  the  true  spirit  of  the  newer  material. 

Thus  it  has  transpired  that  the  McGraw  Building  has  a  floor  con- 
struction which  is  rated  far  below  its  true  safe  carrying  capacity. 
Were  any  floor  loaded  to  failure,  the  latter  would  probably  take  place 
by  shear,  or  rather  diagonal  tension.  In  relation  to  this,  however,  it 
must  be  stated  that  the  speaker  never  has  understood  why  those  in 
actual  charge  of  the  design  of  the  reinforcement  (other  than  the 
author  of  the  paper),  invariably  used  an  odd  number  of  rods  to  resist 
tension.  By  so  doing  it  is  impossible  to  bend  upward  the  same  number 
of  rods  at  each  end  of  a  beam  to  assist  in  resisting  shear,  so-called. 
Thus,  Ave  rods  might  be  used  in  a  given  case  for  tension  reinforce- 
ment. Three  could  be  bent  upward  at  one  end,  but  then  it  is  practicable 
to  bend  up  only  two  at  the  opposite  end  of  the  adjoining  beam  without 
causing  a  congestion  of  steel  over  the  support.  If  the  two  bent  rods 
were  just  sufficient  to  assist  the  shear,  the  three  rods  at  the  other  end 
would  give  50%  better  efficiency  at  that  end;  and  20%  more  resistance 
would  have  been  secured  at  the  weaker  end  by  using  six  rods  of 
smaller  individual  (but  aggregating  the  same  total)  area,  and  bending 
up  three  rods  at  each  end. 

The  floor  forms  were  designed  with  especial  care.  They  were 
collapsible  in  type,  and  were  erected  in  an  exceptionally  substantial 
manner,  so  as  to  be  capable  of  serving  as  platforms  from  which  to 
erect  the  structural  work  of  the  columns.  The  determination,  after- 
ward made,  to  use  the  central  tower  for  erection  purposes,  rendered 
this  special  reason  for  heavy  forms  unnecessary,  but  their  value  was 
repeatedly  shown  for  other  reasons,  and  the  speaker  is  decidedly  of  the 
opinion  that  a  little  extra  material  in  excess  of  that  sometimes  seen, 
is  of  real  economic  advantage.    The  forms  were  designed  by  Mr.  J.  G. 


60  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Goodrich.  Ellendt,  and  were  all  built  in  a  special  shop,  hauled  to  the  building 
site  on  trucks,  and  the  whole  truck-load  hoisted  by  the  central  derrick 
in  a  single  operation  and  set  practically  in  place.  The  speed  actually 
attained  in  erecting  the  building,  shows  how  well  the  form  work  was 
prepared  and  carried  on,  because  that  work  is  the  crucial  part  of  the 
erection  of  all  concrete  work.  Matched  and  dressed  material  was  used 
throughout,  always  well  coated  with  oil,  so  as  to  obviate  the  necessity 
of  special  surface  finish  if  possible.  However,  the  rapid  and  repeated 
use  of  this  material  during  the  winter  soon  disclosed  the  fact,  the 
truth  of  which  has  always  been  held  by  the  speaker,  that  it  would  be 
necessary  to  plaster  the  building,  if  it  was  to  be  given  a  character  on 
a  par  with  the  average  office  structure. 

No  plans  of  the  forms  are  included  in  the  paper,  although  reference 
to  them  appears  at  one  point. 

In  the  speaker's  opinion,  the  tower  used  in  the  erection  of  this 
building  was  really  a  factor  of  large  economy.  For  instance,  all  con- 
crete was  hoisted  to  each  floor  in  buckets  dropped  through  the  elevator 
shafts  to  the  mixers,  which  were  in  the  basement  and  placed  so  as  to 
dump  directly  into  the  buckets  as  they  rested.  The  booms  swung  the 
buckets  so  that  they  could  be  dumped  exactly  at  the  desired  points, 
thus  obviating  the  use  of  other  hoists,  hoppers,  wheel-barrows,  runways, 
etc.  This  method  proved  so  effective  that  very  often  the  cost  of  all 
labor  on  concrete  for  considerable  quantities  would  not  exceed  40  cents 
per  cu.  yd. 

The  speaker  certainly  would  repeat  the  use  of  that  special  con- 
trivance on  a  similar  operation,  except  that  he  would  stiffen  the  struc- 
ture to  a  somewhat  greater  extent,  and  would  use  12  by  12-in.  timbers 
for  corner  posts  instead  of  the  10  by  10-in.  posts  used  in  this  instance. 
The  tower  structure  also  served  as  a  storage  space,  and  was  of  almost 
inestimable  value  in  this  respect,  because  of  the  congested  portion  of 
the  city  in  which  the  building  stands. 

During  cold  weather,  besides  making  use  of  the  salamanders, 
as  described  by  the  author,  the  concrete  was  mixed  with  hot  water  and 
all  aggregates  were  heated  so  as  to  prevent  frozen  lumps  from  getting 
into  the  work.  On  one  operation  with  which  the  speaker  was  con- 
nected he  once  removed  a  lump  of  frozen  sand  from  a  column  in 
which  it  would  have  occupied  about  15%  of  the  total  area.  The 
necessity  of  heating  the  aggregate  is  obvious,  since,  even  when  boiling 
water  is  thrown  into  the  mixer,  it  has  such  speed  of  operation  that  not 
enough  time  elapses  to  thaw  frozen  masses  and  get  them  properly 
distributed,  before  the  mixing  process  is  complete. 

With  the  methods  used  on  the  McGraw  Building,  even  in  the 
coldest  weather,  the  concrete  would  reach  the  point  of  deposit  at  a 
temperature  ranging  from  50  to  75°  fahr.,  and  would  have  attained 
its  initial  set  while  its  temperature  was  still  warm  to  the  touch.     The 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING  61 

salamanders  maintained  a  temperature  in  the  dead  air  spaces  between  Mr.  Goodrich, 
the   beams,   which  often   reached   100°   fahr.,   and   seldom   fell   below 
60°,  even  in  zero  weather. 

Of  course,  the  special  interest  of  this  paper  centers  around  the  type 
of  column.  The  speaker  feels  that,  even  at  the  present  time,  the  de- 
signing of  reinforced  concrete  columns  is  like  working  in  a  darkened 
room,  and  this  is  said  even  after  personally  making  a  large  number  of 
column  tests,  and  after  carefully  analyzing  nearly  a  hundred  others. 
If  carefully  designed,  and  when  a  proper  relation  exists  between  the 
longitudinal  and  spiral  steel,  the  speaker  considers  a  Considere  column 
entirely  practicable  for  a  reasonably  high  building  with  comparatively 
light  floor  loads;  but,  for  more  lofty  structures,  the  opinion  is  gain- 
ing strength,  in  the  speaker's  mind,  that  a  regular  structural  steel 
column  should  be  used,  in  connection  with  reinforced  concrete  girders, 
beams,  and  floors,  if  desired.  This  structural  column,  however,  should 
be  of  some  open  design,  and  it  should  be  completely  filled  with  con- 
crete and  surrounded  by  a  fire-proofing  at  least  3  in.  thick  over  all 
extreme  edges.  Such  a  composite  building  will  be  more  economical 
than  any  other,  in  yearly  carrying  charges,  including  interest  on  first 
cost,  insurance,  maintenance,  heat,  etc.,  and  a  correspondingly  larger 
income  can  be  derived  therefrom. 


62  MEMOIR  OF  NATHANIEL  HENRY  HUTTON  [Memoirs. 

MEMOTES  OF  DECEASED  MEMBERS. 

Note.— Memoirs  will  be  reproducpcl  in  the  volumes  of  Transactions.  Any  information 
which  will  amplify  the  records  as  here  printed,  or  correct  any  errors,  should  be  forwai'ded 
to  the  Secretary  prior  to  the  final  publication. 


NATHANIEL  HENRY  HUTTON,*  M.  Am.  Soc.  C.  E. 


Died  May  8th,  1907. 

Nathaniel  Henry  Hutton  was  born  on  November  16tli,  1833,  in 
Washington,  D.  C,  and  died  in  Baltimore,  Maryland,  on  May  8th, 
1907.  His  earliest  ancestor  in  the  United  States  of  whom  there  is  any 
record,  John  Strangeways  Hutton,  was  born  in  New  York  City  in  1684 
iind  died  in  Philadelphia  in  1792.  His  father  was  James  Hutton,  who 
married  Salome  Rich  of  Boston,  Mass.,  in  Washington,  D.  C. 

Following  the  example  of  his  elder  brother,  the  late  William  Rich 
Hutton,  M.  Am.  Soc.  C.  E.,  "Harry"  (as  he  was  familiarly  called  by 
those  who  knew  him  well)  entered  the  service  of  the  United  States  at 
;ui  early  age,  adopting  the  profession  of  civil  engineering.  Neither  had 
the  advantage  of  a  collegiate  education,  but  they  did  have  the  good 
fortune  to  grow  up  under  the  thorough  training  of  those  days,  in  the 
specially  excellent  schools  of  Alexandria  and  Washington,  taught  by 
men  like  Ben  Hallowell,  Abbot  and  others.  They  made  good  use  of 
those  early  opportunities,  and  by  industry,  faithful  attention  to  duty, 
and  continual  study  of  the  theory  of  engineering  and  the  works  of  able 
engineers,  their  own  experience  and  unusual  natural  talents  enabled 
them  to  pass  through  the  lower  grades  of  the  profession  with  credit  to 
themselves,  and  with  the  respect  and  ever-increasing  confidence  of  their 
superiors  in  their  integrity  and  high  tone,  until  they  had  come  to  rank 
well  among  the  engineers  of  their  period  in  the  special  lines  to  which 
their  attention  was  called. 

Mr.  Button's  work  as  a  surveyor  and  engineer,  up  to  1896,  may  be 
summarized  briefly  as  follows: 

He  was  U.  S.  Assistant  Engineer  on  explorations  and  surveys  for 
the  Pacific  Railroad  west  of  the  Missouri  River,  on  the  32d  and  35th 
Parallels,  from  1853  to  1856,  inclusive;  Chief  Engineer  of  the  El  Paso 
and  Fort  Yuma  wagon  road  (Department  of  Interior)  during  1857  and 
1858;  Surveyor  on  the  western  boundary  of  Minnesota  (Department  of 
Interior)  during  1859  and  1860;  U.  S.  Assistant  Engineer  on  the  de- 
fenses of  Baltimore  from  1861  to  1865 ;  U.  S.  Assistant  Engineer  in 
charge  of  the  improvement  of  the  Patapsco  River  from  1867  to  1876, 
nnd  on  the  Western  division  of  the  Virginia  Central  Water  Line  (sur- 
vey 1874  to  1875)  ;  and  from  1876  until  his  death  he  was  Engineer  to  the 
Harbor  Board  of  Baltimore;  he  was  also  U.  S.  Assistant  Engineer  in 
charge  of  surveys  for  a  ship  canal  to  connect  the  Chesapeake  and  Dela- 
*  Memoir  prepared  by  William  P.  Craighill,  Past -President,  Am.  Soc.  C.  K. 


Memoirs.]  MEMOIR  OF  NATHANIEL  HENRY  HUTTON  63 

ware  Bays  during  1878  and  1879;  Consulting  Engineer  for  a  project  for 
a  ship  canal  between  Philadelphia  and  the  Atlantic  Ocean  in  1894  and 
1895;  and  Consulting  Engineer  for  a  projected  ship  canal  to  connect 
Lake  Erie  and  the  Ohio  Eiver  in  1895  and  1896. 

For  many  years  previous  to  1896,  and  up  to  the  time  of  his  death, 
Mr.  Hutton  had  been  Chief  Engineer  to  the  Harbor  Board  of  the  City 
of  Baltimore.  That  he  held  this  office  so  many  years,  during  the  admin- 
istrations of  mayors  and  councils  of  opposing  political  parties,  is  proof 
that  his  services  were  considered  so  valuable  as  to  be  almost  indispen- 
sable. Later,  he  became  President  of  the  Harbor  Board,  as  well  as 
Chief  Engineer. 

The  following  tribute  from  the  Harbor  Board  shows  the  high  esteem 
in  which  he  was  held  by  his  associates,  and  it  may  be  said  with  truth 
that  this  was  the  sentiment  of  the  business  men  of  Baltimore  who  were 
best  acquainted  with  his  work  and  ability: 

"The  death  of  Major  Nathaniel  H.  Hutton,  Engineer  of  the  Harbor 
Board  of  Baltimore  City,  comes  at  a  time  and  under  conditions  which 
cause  especially  deep  feelings  of  sorrow  and  regret  in  the  minds  of  the 
members  of  the  Harbor  Board. 

"Immediately  after  the  fire  of  February  7th  and  8th,  1904,  he  was 
called  upon  by  the  citizens  of  Baltimore  to  suggest  and  design  plans  for 
the  new  docks  and  the  improvements  of  the  harbor  of  this  City.  The 
preparations  of  these  plans,  together  with  his  other  duties  as  engineer 
of  the  Harbor  Board,  devolved  upon  him  a  very  great  amount  of  skill- 
ful professional  work,  and  it  is  probable  that  he  unconsciously  over- 
taxed his  strength  in  this  way. 

"The  influence  which  Major  Hutton  has  exerted  upon  the  plans  for 
the  improvement  of  the  Harbor,  cannot  be  estimated.  He  has  not  lived 
to  see  the  realization  of  what  he  has  planned,  bvit  there  can  be  no  doubt 
that  his  activity  and  experience  in  this  great  work  will  be  appreciated 
by  his  successors,  and  the  citizens  of  Baltimore,  when  the  full  effects 
of  his  labors  and  efforts  are  realized. 

"Major  Hutton  was  an  engineer  of  rare  ability  and  of  vast  and 
varied  experience.  He  was  a  gentleman  of  the  old  school,  and  a  most 
faithful  engineer  and  honest  public  servant. 

''Resolved,  that  in  the  death  of  Major  Nathaniel  H.  Hutton,  the 
City  of  Baltimore  has  been  deprived  of  a  noble  and  trusted  citizen  and 
a  capable  and  conscientious  public  servant,  who  has  devoted  many  years 
of  his  life  to  her  interests. 

"Resolved,  that  the  members  of  the  Harbor  Board,  who  particularly 
appreciate  the  full  measure  of  loss  suffered  by  his  death,  tender  their 
sympathies  to  the  family  of  the  deceased,  and  that  these  Resolutions  be 
spread  upon  the  Minutes  of  the  Board." 

There  are  also  appended  resolutions  adopted  May  19th,  1907,  by  the 
Board  of  Public  Improvements,  of  which  Mr.  Hutton  was  a  prominent 
member : 

"At  a  special  meeting  of  the  Board  of  Public  Improvements  held 
this  date  called  to  take  action  on  the  death  of  Major  N.  H.  Hutton, 


64  MEMOIR  OF  NATHANIEL  HENRY  HUTTON  [Memoirs. 

President  and  Chief  Engineer  of  the  Harbor  Board,  the  following  reso- 
lutions were  adopted : 

"Resolved,  that  by  the  death  of  Major  Button  the  City  of  Baltimore 
has  lost  a  most  faithful  and  efficient  public  officer,  whose  long  service 
as  Harbor  Engineer  here  and  extended  experience  on  important  public 
works  elsewhere  made  his  services  invaluable  to  this  city. 

"Also  by  his  death,  we,  his  fellow  members  of  the  Board  of  Public 
Improvements,  have  lost  a  trusted  friend  and  wise  counsellor,  whose 
uniformly  genial  and  courteous  nature  greatly  endeared  him  to  us. 

"We  extend  to  his  family  our  sincere  and  heartfelt  sympathy  in 
their  great  sorrow." 

Mr.  Hutton  was  a  Charter  Member  and  Vice-President  of  the  Engi- 
neers' Club  of  Baltimore.  At  his  death  the  Club  took  the  following 
action  in  his  honor: 

'^Whereas,  We,  the  members  of  the  Engineers'  Club  of  Baltimore, 
have  learned  with  sincere  sorrow  of  the  death  of  our  fellow  member. 
Major  N.  H.  Hutton;  and  whereas  we  recognize  his  earnest  efforts,  as 
a  Charter  Member  and  Vice-President,  to  promote  the  welfare  of  the 
Club,  and  the  active,  friendly  and  generous  interest,  manifested  by 
him,  in  establishing  its  success : 

"Resolved,  that  in  his  death  the  Engineers  Club  of  Baltimore  has 
been  deprived  of  a  distinguished  member  and  a  Loyal  and  Honoured 
Friend." 

Mr.  Hutton  was  also  an  architect  of  decided  ability,  as  is  shown  by 
the  outcome  of  the  designs  proposed  by  the  firm  of  Hutton  and  Mur- 
dock,  of  which  he  was  a  member  for  several  years,  for  the  construction 
and  alteration  of  a  number  of  churches,  dwelling-houses  and  warehouses 
in  Baltimore,  Washington,  Virginia  and  Pennsylvania.  One  of  his 
designs  for  a  highway  bridge  in  Baltimore  was  considered  by  a  very 
judicious  board  to  be  the  best  among  five  that  were  submitted.  Not 
only  was  Mr.  Hutton  esteemed  as  an  able  engineer  and  architect  and  a 
capable  and  faithful  official,  but  he  was  admired  and  loved  by  his 
friends  in  an  unusual  degree.  A  few  extracts  are  appended  from  many 
testimonials  that  have  been  received  as  proof  of  the  statements  already 
made. 

After  a  long  intercourse,  under  conditions  which  often  test  men's 
character,  long-drawn-out  surveys  among  the  rough  surroundings  of 
camp  life,  in  the  midst  of  Indians  and  uncultivated  and  often  lawless 
frontier  people,  both  male  and  female,  one  of  his  closest  friends  writes : 

"  'Tis  said  that  you  must  sleep  with  a  man  to  learn  his  peculiarities. 
Well,  if  this  is  true,  Harry  and  I  ought  to  have  become  pretty  well 
acquainted,  for  the  nights  we  stretched  ourselves  on  the  ground  under 
the  same  blanket,  ate  oiir  grub  out  of  the  same  tin  pan,  and  drank  our 
coffee  out  of  the  same  tin  cup,  ran  through  years,  and  during  the  entire 
time  our  affection  became  closer.  It  was  only  necessary  to  know  him  to 
love  him,  and,  of  the  many  acquaintances  I  have  made  during  a  long 
and  varied  life,  I  have  yet  to  meet  the  man  who  excelled  him  in  the 


Memoirs,]  MEMOIR  OF  NATHANIEL  HENRY  HUTTON  66 

noble  qualities  of  head  and  heart  which  he  possessed.  He  was  one  of 
Nature's  noblemen,  a  conscientious  Christian  whose  only  fear,  if  he 
knew  what  fear  was,  was  to  do  wrong,  and  whose  sense  of  honor  was  as 
firmly  fixed  as  the  everlasting  hills." 

Another,  with  whom  Mr.  Hutton  had  close  professional  and  personal 
contact  in  Baltimore,  gives  the  following  high  testimonial  from  himself 
and  others  of  their  mutual  associate: 

"All  of  us  had  the  highest  appreciation  of  his  ability  as  an  engineer 
and  of  the  value  of  his  services  to  the  city.  He  had  been  our  Harbor 
Engineer  for  so  many  years  that  he  had  become  indispensable  in  the 
working  of  our  city  government.  His  advice  was  frequently  sought  by 
municipal  engineers  and  other  municipal  officials,  and  his  opinion  was 
always  respected  on  all  engineering  questions.  He  was  progressive, 
broad  and  liberal  in  his  views,  yet  conservative  enough  to  hold  down 
some  of  us  younger  and  rasher  engineers.  He  was  a  conciliating  and 
harmonizing  influence  at  all  gatherings  of  engineers  and  meetings  of 
boards  and  commissions.  His  personality  was  such,  and  his  manner 
was  so  genial  and  kindly,  that  he  could  regulate  or  harmonize  where 
others  could  not,  and  yet  always  retain  the  regard  and  affection  of  his 
associates. 

"Because  of  his  years  of  experience  and  of  his  broad  learning,  his 
place  in  our  municipal  government  will  be  hard  to  fill.  His  place  in 
our  affections  can  never  be  filled." 

Another  who  had  served  with  Mr.  Hutton  very  closely  for  many 
years  adds: 

"As  an  engineer,  he  was  capable,  careful,  eminent  and  prominent, 
and  was  consulted  in  the  development  of  many  projects  of  National 
importance.  On  undertaking  any  new  work  he  sought  the  results  and 
opiniond  of  others  of  distinction  and  after  giving  careful  consideration 
formulated  his  plans. 

"As  a  public  official,  he  was  earnest,  honest  and  faithful,  possessing 
a  keen  power  of  penetration,  and  his  approval  always  carried  weight. 

"As  a  man,  he  was  modest  and  retiring,  affable  and  lovable,  with 
ever  a  kind  word  for  his  fellow-man,  be  he  high  or  low,  and  all  in  all 
a  splendid  type  of  a  gentleman." 

Still  another  says: 

"I  was  thrown  in  intimate  relations  with  him.  He  was  always  to  me 
the  embodiment  of  a  true  gentleman,  in  the  highest  and  best  sense  of 
that  word;  honorable  and  truthful,  above  suspicion,  always  courteous 
and  always  manly. 

"As  an  engineer,  he  was  well  trained  and  on  broad  lines.  I  had 
great  confidence  in  him,  and  frequently  consulted  him  about  difficult 
problems  coming  up  in  my  work,  and  always  got  sound  and  helpful 
advice.  If  I  were  called  upon  to  name  some  special  characteristic  of 
Major  Hutton,  which  distinguished  him  as  an  engineer,  I  should  say 
that  good  judgment  was  his  strong  point. 

"His  death  leaves  a  great  blank,  both  professionally  and  socially. 
My  feelings  for  Major  Hutton  were  those  of  real,  genuine  affection. 


66  MEMOIR  OF  NATHANIEL  HENRY  HUTTON  [Memoirs. 

and  I  believe  that  most  men  who  came  in  close  contact  with  him  had 
the  same.  It  is  difficult  to  imagine  a  true  man  having  any  sentiments 
for  Major  Hutton  other  than  those  of  the  profoundest  confidence  and 
respect." 

The  writer  knew  Mr.  Hutton  for  more  than  forty  years,  both  profes- 
sionally and  socially,  and  can  fully  bear  testimony  to  the  fact  that  what 
is  said  by  others  in  what  precedes  is  not  exaggerated.  His  domestic 
life  was  charming  and  lovely. 

In  early  manhood,  Mr.  Hutton  married  Miss  Meta  Van  Ness, 
daughter  of  Colonel  Eugene  Van  Ness  of  the  United  States  Army, 
who  was  a  member  of  the  well-known  and  distinguished  family  of  that 
name  in  the  State  of  New  York.  One  of  Mrs.  Hutton's  ancestors  was 
Admiral  Van  Ness  of  Holland,  who  lived  in  1653 ;  and  in  Scotland  her 
lineage  dates  distinctly  and  honorably  at  least  to  1542. 

Mr.  Hutton  passed  from  time  to  eternity  in  May,  1907,  and  his 
devoted  wife  followed  in  September.  They  left  three  children,  all  resi- 
dent in  Baltimore,  Mr.  Harry  Hutton,  Mrs.  S.  S.  Busby  and  Mrs.  C. 
H.  Wyatt. 

Mr.  Hutton  was  elected  a  Member  of  the  American  Society  of  Civil 
Engineers  on  June  3d,  1896. 


AMERICAN    SOCIETY 


OF 


CIVIL  ENGINEERS 


February,    1908. 


PROCEEDINGS  =  VOL.  XXXIV— No.  2 


AMERICAN 
SOCIETY  OF 

CIVIL 
^ENGlNEEpSi 


By 
W.LUAM  P.  Morse 


Published  at  the  House  of  the  Society,  220  West  Fifty-seventh  Street,  New  York 
the  Fourth  Wednesday  of  each  Month,  Except  June  and  July. 

Copyrighted  1908,  by  the  American  Society  of  Civil  Engineers. 

Entered  as  Second-Class  Matter  at  the  New  York  City  Post  Office,  December  15th,  1896. 

Subscription,  $6  per  annuna. 


Vol.    XXXIV.  FEBRUARY,  1908.  No.  2. 


AMEKIOAN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED    185  3. 


PAPERS   AND   DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


CONTENTS. 

Papers :  page. 

The  Electrification  of  the  Suburban  Zone  of  the  New  York  Central  and  Hudson 
River  Railroad  in  tne  Vicinity  of  New  York  City. 

By  William  J.  WiLGUs,  BI.  Am.  Soc.  C.  E 68 

The  Flood  of  March.  1907,  in  the  Sacramento  and  San  Joaquin   River  Basins, 
California. 
By  W.  B  Clapp,  M.  Am.  Soc.  C.  E.,  E.  C.  Murphy,  Assoc.  M.  Am.  Soc.  C.  E., 
AND  W.  F.  Martin,  Jun.  Am.  Soc.  C.  E  99 

Discussions : 

The  Reinforced  Concrete  Work  of  the  McGraw  Building. 

By  Messrs.  T.  L.  Condron  and  F.  F.  Sinks,  E.  AV.  Stern,  L.  J.  Mensch.  and 
P,  E.  Stevens 149 

The  Use  of  Reinforced  Concrete  in  Engineering  Structures  :   An  Informal  Discus- 
sion. 
By  Messrs.  E.  P.  Goodrich,  Edwin  Thacher.  Sanford  E.  Thompson,  W.  H. 
Burr,  T.  Kennard  Thomson,  D.  W.  Krellwitz,  Guy  B.  Waite,  and  C.  L. 
Slocum 169 

Memoir : 

Charles  Paine,  Past-President,  Am.  Soc.  C.  E 196 

PLATES. 

Plate  V.              Map  of  the  Electric  Zone,  New  York  Central  and  Hudson  River  Rail- 
road, in  the  Vicinity  of  New  York  City 69 

Plate  VI.            Port  Morris  and  Yonkers  Power  Stations,  Cable  Tower,  Etc 71 

Plate  VII.          Cross-Section  and  Plan  of  Port  Morris  Power  Station 73 

Plate  VIII.        Aerial  and  Duct  Tran^  mission  Lines.   75 

Plate  IX.            Splicing  (Jhambsr,  Submarine  Crossing,  and  Sub-Station 75 

Plate  X.  Third-rail,  Circuit- Breaker  House,  Operation  of  Third-rail  in  Winier, 

and  Electric  Locomotive 75 

Plate  XI.            Electri3  Trains,  "  Jump?r,"  and  Reactance  Bond 78 

Plate  XII.          Electric  Locomotive,  and  Bi- polar  Gearless  :Vlotor 79 

Plate  XIII.        Train  Braking  Charts  for  Distance  and  Speed 81 

Plate  XIV.         Terminals  at  North  White  Plains,  Hi^h  Bridge,  and  Wakefield , .  83 

Plate  XV.          Harmon  Shops,  and  Grand  Central  Yard 85 

Plate  XVI.        Present  Grand  Cential   Station  and  Train-shed,  and   Excavation    in 

Progress  at  <Grand  Central  Terminal 87 

Plate  XVII.       Contrast  Between  Smoke  Conditions  at  Grand  Central  Terminal,  Be- 
fore and  After  Electrifleation 89 

Plate  XVIII.     Proposed  Bronx  Station  and  Improvement,  and  Typical  Grade  Cross- 
ing Elimination      91 

Plate  XIX.         Grade  Crossing  Elimination,  Marble  Hill  Cut-off,  Signal  Tower,  etc. ...  93 

Plate  XX.           Four  and  Six-tracking,  Typical  Signals,  and  Typical  Fencing 93 

Plate  XXI.         Table  3:  Comparative  Tests  of  Steam  and  Electric  Locomotives 95 

Plate  XXII.       Profile  from  Grand  Central  Station  to  White  Plains 97 

Plate  XXIII.     Map    of    I>rainage    Basins  of    Sacramento  and   San  Joaquin  Rivers, 

California 105 

Plate  XXIV.      Reinforcement  and  Concreting.  Watson  Building 151 

PlateXXV.       Maniitacturets'  Furniture  F-xehange  Building 153 

Plate  XXVI.     Reinforcement  and  Concreting,  Manufacturers'  Furniture  Exchange 

Building 1 55 

Plate  XXVII.    Exterior  and  Interior,  Manufacturers'  Furniture  Exchange  Building. .  157 


Vol.  XXXIV.  FEBRUARY,  1908.  No.    2. 


AMEEICAN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED     1852. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


THE  ELECTRIFICATION  OF  THE  SUBURBAN  ZONE 

OF  THE 
NEW  YORK  CENTRAL  AND  HUDSON  RIVER  RAIL- 
ROAD IN  THE  VICINITY  OF  NEW  tORK  CITY. 


By  William  J.  Wilgus,  M.  Am.  Soc.  C.  E. 
To  BE  Presented  March  18th,  1908. 


The  recent  successful  completion  of  the  electrification  of  the  service 
of  the  New  York  Central  and  Hudson  River  Railroad  entering  the 
Grand  Central  Terminal,  New  York  City,  marks  such  an  important  step 
in  the  progress  of  the  art  of  transportation  that  a  paper  seems  at  this 
time  appropriate,  explaining  the  reasons  for  the  abandonment  of  steam, 
the  general  features  of  construction  and  operation,  and  the  results. 

Two  decades  have  passed  since  electricity  in  the  United  States  first 
commenced  its  important  career  in  the  field  of  lighter  traffic;  but  only 
within  the  past  few  months  has  it  fairly  met  its  steam  rival  in  heavy- 
traction  trunk-line  service. 

Reasons  for  Delay  in  Electrification  of  Trunh  Lines. — The  reason 
for  this  delay  is  not  far  to  seek.  The  steam  locomotive,  during  its 
lifetime  of  eighty  years,  has  been  developed  into  a  wonderfully  reli- 
able, efficient,  and  powerful  machine,  deep-seated  in  thp  affections  of 

Note. — These  papers  are  issued  before  the  date  set  for  presentation  and  discussion. 
Correspondence  is  Invited  from  those  who  cannot  be  present  at  the  meeting,  and  may  bo 
sent  by  mail  to  the  Secretary.  Discussion,  either  oral  or  written,  will  be  published 
in  a  subsequent  number  of  Prorerdinas.  and,  when  finally  closed,  the  papers,  with 
discussion  in  full,  will  be  published  In  Transactions, 


PLATE  V. 

PAPERS  AM.  80C.  C.   E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  4  H.  R.  R.  R. 


Papers.]   ELECTKII'ICATION,  SUBUltBAN  ZONE,  N.  Y.  C.  &  11.  K.  R.  R.        G9 

the  railroad  world.  With  the  conservatism  naturally  born  of  these 
conditions  is  the  reluctance  of  stockholders  to  spend  vast  sums  for 
changes  of  unproven  financial  value. 

There  is  no  cause  for  surprise,  therefore,  that  electricity,  so  com- 
monly associated  in  the  mind  of  the  railroad  officer  with  light  street- 
car traffic,  has  not  been  seriously  considered  as  a  substitute  for  steam, 
until  special  problems  have  arisen  demanding  some  escape  from  the 
limitations  and  nuisances  incident  to  the  use  of  steam  locomotives. 

The  very  fact  that  steam  locomotives  have  grown  so  in  size  and 
power  makes  them  more  objectionable  as  emitters  of  increased  volumes 
of  noise,  smoke,  gas,  and  cinders. 

The  first  important  instance  of  the  use  of  electricity  on  a  large 
scale  was  in  utilizing  electric  locomotives  to  push  solid  trains,  with 
their  inactive  steam  locomotives,  through  the  Baltimore  Tunnel  of  the 
Baltimore  and  Ohio  Railroad.  In  this  instance,  electricity  was  adopted 
as  an  aid,  not  as  a  substitute  for  steam. 

Reasons  for  Electrification  of  New  York  Central. — As  early  as  1899, 
thought  was  given  to  the  use,  on  the  New  York  Central,  of  electricity 
for  curing  the  evils  at  the  entrance  to  the  Grand  Central  Terminal; 
but  it  was  not  until  1903  that  the  objectionable  atmospheric  conditions 
in  the  Park  Avenue  Tunnel,  and  the  congestion  of  traffic  at  the  termi- 
nal, precipitated  legislative  action  directing  the  complete  abandonment 
of  the  steam  locomotive  in  Park  Avenue  south  of  the  Harlem  River, 
within  a  period  of  five  years  terminating  July  1st,  1908. 

In  the  same  year  the  railroad  company  and  the  city  agreed  upon 
radical  changes  at  the  terminal,  which  were  possible  only  with  the 
abandonment  of  steam.  From  a  civic  standpoint,  the  most  important 
of  these  changes  is  the  depression  of  the  whole  terminal,  so  as  to 
permit  the  extension  of  highways  over  the  tracks  from  Forty-fifth  to 
Fifty-sixth  Streets,  inclusive,  and  the  continuation  of  Park  Avenue,  140 
ft.  wide,  within  the  same  limits,  thus  joining  two  sections  of  the  city 
hitherto  separated  for  |  mile  by  an  impassable  barrier  of  railroad  yards 
and  structures. 

Reasons  for  Extended  Scope  of  Electrification. — A  careful  analysis 
of  the  situation  soon  proved  the  absurdity  of  terminating  the  electric 
zone  at  or  near  the  Harlem  River. 

Immediately  north  of  that  point  is  Mott  Haven  Junction,  where 
the  line  splits,  one  leg  known  as  the  Harlem  Division  continuing  north 


70       ELECTRIFICATION,  SUBUEBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.   [Papers. 

to  Chatham  on  the  Boston  and  Albany  Railroad,  and  the  other  consti- 
tuting the  main  line  of  the  Hudson  Division,  bearing  to  the  west  and 
north,  along  the  banks  of  the  Hudson  River,  to  Albany  and  beyond. 
At  Woodlawn  Junction,  on  the  Harlem  Division,  is  the  point  of  conflu- 
ence with  the  New  York,  New  Haven  and  Hartford  Railroad,  the  very 
large  passenger  traffic  of  which  flows  over  the  rails  of  the  New  York 
Central  to  and  from  the  Grand  Central  Terminal,  a  distance  of  12  miles, 

From  the  Grand  Central  Terminal  to  Woodlawn  Junction,  the 
Harlem  Division  is  four-tracked,  but  for  the  remainder  of  the  distance 
within  the  territory  under  discussion,  but  two  tracks  existed  for  hand- 
ling all  classes  of  traffic.  Similarly,  on  the  main  line,  from  the  junction 
at  Mott  Haven,  two  tracks  w'ere  called  upon  to  transport  both  passenger 
and  freight  trains,  except  on  the  section  between  Spuyten  Duyvil  and 
Scarborough,  where  a  third  track  aided  to  some  extent. 

In  addition  to  the  extremely  heavy  through  passenger  train  service 
from  the  New  England,  northern  and  western  States  of  the  Union, 
and  Canada,  there  is  an  important  local  traffic  extending  as  far  out  as 
Harmon,  on  the  Hudson  Division,  a  distane  of  33  miles.  North  White 
Plains,  on  the  Harlem  Division,  a  distance  of  24  miles,  and  Stamford, 
on  the  New  York,  New  Haven  and  Hartford  Railroad,  a  distance  of  34 
miles,  from  the  Grand  Central  Station. 

A  further  burden  on  the  four-track  stem  between  the  terminal  and 
Mott  Haven  Junction  is  the  hauling  between  those  points  of  "dead" 
equipment,  because  of  inadequate  storage  space  at  the  station. 

From  this  recital  it  will  be  seen  that  a  termination  of  the  electric 
zone  at  the  Harlem  River,  or  at  Mott  Haven  Junction  just  above  the 
river,  would  entail  the  stoppage  a;id  change  of  motive  power  from 
steam  to  electricity  and  vice  versa,  of  all  kinds  of  traffic,  at  a  point 
peculiarly  subject  to  congestion.  Moreover,  the  physical  conditions  in 
the  neighborhood  precluded  the  construction  of  the  necessary  facilities 
for  the  storage  and  care  of  motive  power. 

Because  of  these  fundamental  objections,  and,  moreover,  guided  by 
the  broad-minded  policy  that  growth  of  traffic  responds  to  the  use  of 
electricity,  the  company  decided  to  extend  the  limits  of  the  electric 
zone  to  the  northerly  termini  of  the  suburban  territory,  at  Harmon 
and  North  White  Plains,  where  ample  space  is  available  for  loops, 
yard  tracks,  and  buildings.  The  geography  of  the  territory  is  shown 
on  Plate  V. 


PLATE  VI. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARv,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  &  H.  R.  R.  R. 


Papers.]   ELECTKIiaCATION,  SUHUJtBA]^  ZONK,  N.  Y.  C.  &  H.  R.  R.  R.        71 

Reasons  for  Other  Improvements. — This  decision,  and  the  demands 
of  growing  traffic  for  more  and  better  facilities,  led  to  the  adoption  of 
plans  for  a  new  Grand  Central  Station  with  two  track  levels;  the 
separation  of  track  grades  and  a  new  overhead  eight-track  station  at 
Mott  Haven;  the  elimination  of  all  grade,  street,  and  highway  crossings; 
the  "four-tracking"  of  both  divisions  as  far  as  the  termini;  many  new 
and  enlarged  passenger  and  freight  stations;  new  electric  automatic 
signals,  and  electric  interlocking  plants ;  and  many  important  revisions 
of  alignment  and  grades. 

Reasons  for  Adopting  the  Direct-Current  System. — About  this  time 
the  battle  had  just  opened  in  the  United  States  between  the  two  rival 
systems  of  electricity — direct  and  alternating  current.  Of  coiirse,  the 
advocates  of  each  argued  that  the  pther  was  unsuited  to  New  York 
Central  conditions,  and  it  was  "only  after  lengthy  and  thorough  consid- 
eration that  the  direct-current  system  was  selected. 

The  principal  reasons  for  this  conclusion,  apart  from  technical 
points,  may  be  summarized  as  insufficient  practical  development  of  the 
alternating-current  system  for  a  trunk-line  problem  reqiiiring  absolute 
reliability  of  service,  restricted  clearances  which  forbade  the  use  of 
overhead  conductors,  and  legal  obstacles  to  the  use  of  overhead  trolley 
wires  carrying  high  voltages  within  the  limits  of  the  City  of  New  York. 

Reasons  for  Not  Using  Alternating-Current  Equipment  on  a  Direct- 
Current  System. — Some  time  after  this  decision  had  been  made,  and 
apparatus  had  been  ordered,  the  company  was  urged  by  outside  inter- 
ests to  abandon  the  tyi^e  of  equipment  suited  exclusively  to  the  direct- 
current  system,  and  adopt  another  type  which  could  operate  on  both 
direct  and  alternating  currents.  It  was  claimed  that,  by  making  this 
change,  the  equipment  would  be  available  for  use  on  later  extensions 
of  the  electric  zone  where  there  were  no  physical  or  legal  objections 
to  the  use  of  alternating  current.  The  wisdom  of  adhering  to  the  type 
of  equipment  already  chosen  has  been  proven  by  recent  comparative 
tests  of  locomotives  of  the  two  types  under  exactly  the  same  conditions, 
which  demonstrate  that  the  one  designed  only  for  direct  current  con- 
sumes from  15  to  25%  less  current  than  the  one  intended  for  use  on 
both  systems.  This  will  effect  a  saving  to  the  company  of  at  least 
$140  000  per  annum.  If  to  this  item  is  added  the  economy  resulting 
from  less  locomotive  ton-miles  per  annum  because  of  the  lower  weights 
of  locomotive  per  unit  of  capacity,  and  lower  wages,  fixed  charges,  and 


72       ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.  [Papers. 

maintenance  of  equipment,  because  of  the  smaller  number  needed  to 
do  the  same  work,  the  total  saving  for  the  ultimate  electric  zone,  result- 
ing from  adherence  to  the  adopted  type  of  direct-current  locomotive, 
will  be  approximately  $300  000  per  annum. 

Reasons  for  Duplicate  Power  Stations  and  Transmission  Lines. — 
One  of  the  strongest  arguments  advanced  against  the  substitution  of 
electricity  for  the  well-tried  steam  locomotive,  for  the  movement  of 
the  most  important  passenger,  mail,  and  express  service  in  the  country, 
is  the  vulnerability  of  power  stations  and  distributing  systems  to  failures 
of  the  class  which  affect,  not  one,  but  all,  units.  To  overcome  this  well- 
founded  criticism,  two  cross-connected  power  stations  were  decided 
upon,  accessible  to  both  rail  and  boat  coal;  and  each  with  sufficient 
capacity,  utilizing  its  spare  unit,  and  working  "overload,"  to  carry  the 
entire  demand  of  the  service  at  the  rush  hours  shovild  the  other  fail. 
It  was  considered  that  the  growing  familiarity  of  the  operating  force 
with  the  new  conditions,  and  the  elimination,  in  time,  of  unsuspected 
defects  of  installation,  would  later  make  the  surplus  capacity  available 
for  other  uses,  such  as  increased  demands  of  traffic,  the  movement  of 
freight  trains  by  electricity,  and  the  operation  of  the  terminals  of  the 
company  on  the  west  side  of  Manhattan  Island.  As  a  further  precau- 
tion, duplicate  transmission  lines  were  adopted  in  the  more  important 
portions  of  the  territory,  so  that  the  failure  of  one  would  still  leave 
the  other  effective  for  the  uninterrupted  movement  of  trains. 

Reasons  for  Storage  Batteries. — Even  with  these  two  safeguards, 
there  appeared  to  be  vulnerable  places,  where  accidents  might  put 
essential  features  of  the  service  out  of  commission,  and  to  overcome 
this,  as  well  as  to  make  suitable  regulation  of  violent  fluctuations  of 
load  on  the  power  stations  and  sub-stations,  storage  batteries  were 
adopted  with  capacity  sufficient  to  tide  over  the  usual  maximum  periods 
of  interruption  of  current  supply,  that  experience  elsewhere  has  shown 
may  be  expected. 

Reasons  for  Combined  Locomotive  and  Multiple-V nit  Practice. — 
While,  necessarily,  through  trains  with  cars  originating  at  far  distant 
points  must  be  hauled  by  electric  locomotives  within  the  electrified 
territory,  it  was  evident  from  the  start  that,  for  the  company  to  reap 
the  full  advantage  from  its  expenditures,  the  multiple-unit  type  of 
suburban  equipment  should  be  adopted  that  elsewhere  had  been  shown 
was  essential  for  the  propagation  of  traffic,  and  the  simplification  of 


PLATE  VII. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,   1908. 

WILQUS    ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  0.  &  H.  R.  R.  R. 


^-m,.^ 


^_  fj^.-l:  -a.  It        IF       f- 


Fig.  1.— Typical  Cross-Section  of  the  Port  Morris  Power  Station. 


Fig.  2.— Typical  Plan  of  Port  Morris  Power  Station. 


Papers.]  ELECTEIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.       73 

operation  in  congested  terminals.  By  dispensing  with  locomotives  in 
suburban  service,  and  equipping  the  individual  cars  with  electric 
motors  controlled  from  either  end  of  the  train,  it  becomes  possible  to 
meet  the  demand  of  the  public  for  less  interval  between  trains,  and  at 
the  same  time  regulate  the  cost  of  operation  to  the  volume  of  traffic  at 
various  periods  of  the  day.  The  absence  of  locomotives,  and  the  distri- 
bution of  power  among  the  cars  practically  eliminates  switching,  and 
movements  to  and  from  engine-houses,  with  a  resultant  great  reduction 
of  the  causes  that  congest  terminals.  A  twofold  character  of  equip- 
ment, therefore,  was  adopted — locomotives  for  through  trains,  and  mul- 
tiple-unit cars  for  the  passenger  service  confined  to  the  electric  zone. 

Awarding  First  Contracts. — With  all  the  foregoing  questions  settled, 
plans  and  specifications  were  actively  prepared,  and  contracts  awarded 
in  the  fall  of  1903  for  the  apparatus  requiring  the  longest  time  for 
delivery,  including  power-station  machinery  and  locomotives.  Later, 
arrangements  were  made  for  the  remaining  items  of  the  installation, 
either  by  contract  or  by  company  forces. 

General  Features  of  Construction. 

Principal  Elements  of  the  Installation. — The  principal  elements  of 
the  installation  are  the  duplicate  power  stations  for  generating  3-phase, 
11  000-volt,  25-cycle  alternating  current ;  the  high-tension  transmission 
lines  for  distributing  this  current  to  the  sub-stations;  the  sub-stations 
for  transforming  and  converting  the  high-tension  alternating  current 
to  660-volt  direct  current,  and  for  the  storage  of  current  in  batteries; 
the  direct-current  transmission  lines  for  the  distribution  of  energy 
to  the  working  conductors;  the  third-rail  and  overhead  conductors  at 
special  places,  known  as  working  conductors,  for  the  delivery  of  the 
660-volt  current  to  the  contact  shoes  on  locomotives  and  cars;  the 
electrical  equipment ;  repair  shops  and  inspection  facilities ;  interchange 
terminals  for  electric  and  steam  power;  and  last,  but  not  least,  the 
building  up  of  an  operating  organization  to  make  all  this  intricate 
machinery  a  working  success. 

Power  Stations. — Each  power  station  is  equipped  initially  with  six- 
teen water-tube,  625-h.p.  boilers,  with  superheaters  and  mechanical 
stokers,  and  four  5  000-k.w.  Curtis  turbo-generators,  together  with  the 
necessary  condensers,  pumps,  exciters,  feed-water  heaters,  and  appur- 
tenances.    Additional  space  is  provided  in  the  buildings,  for  a  later 


74       ELECTRIFICATION,  SUBURBAN  ZONE,  N.  T.  C.  &  H.  R.  R.  R.  [Papers. 

expansion  of  capacity  to  the  extent  of  50%  of  the  initial  installation. 
It  will  thus  be  seen  that  each  station  has  a  present  normal  capacity  of 
approximately  28  000  h.p.  (20  000  k.w.),  with  provision  for  an  ultimate 
increase  to  approximately  42  000  h.p.  (30  000  k.w.),  or  a  combined 
ultimate  normal  capacity  of  approximately  84  000  h.p.  (60  000  k.w.). 
The  two  stations  are  electrically  cross-connected,  so  that,  for  all  prac- 
tical purposes,  they  act  as  one. 

Both. power  stations  are  supplied  with  mechanical  plants  for  trans- 
ferring coal  from  car  or  boat  to  overhead  bins,  each  station  having  a 
storage  capacity  of  3  500  tons,  equal  to  9  days'  supply  under  maximum 
conditions. 

A  pilot  switch-board  is  located  in  the  gallery  of  each  power  station, 
but  the  important  control  apparatus,  including  oil  switches,  is  placed 
in  a  separate  building,  so  that  serious  trouble  in  the  main  structure 
will  not  disable  or  injure  what  may  be  termed  the  brains  of  the  system. 

The  buildings  are  constructed  substantially,  of  concrete,  brick  and 
steel,  on  stable  foundations,  and  with  an  architectural  treatment  suited 
to  the  purposes  for  which  they  are  designed.  The  twin  stacks  at  each 
station  are  of  perforated  radial  brick,  have  an  average  internal  diam- 
eter of  16  ft.  3  in.,  and  rise  to  a  height  of  267  ft.  above  the  ground. 

A  noteworthy  fact  may  be  recorded  that  illustrates  one  of  the 
advantages  of  turbo-generators  in  the  economical  design  of  power- 
station  buildings.  The  capacity  required  at  the  Yonkers  station  is 
only  110  cu.  ft.,  and  that  at  the  Port  Morris  station  115  cu.  ft.  per 
k.w.,  as  compared  with  from  170  to  255  cu.  ft.  at  the  more  important 
reciprocating-engine  plants  in  New  York  City. 

The  maximum  calculated  4-min.  peak  load  on  both  stations  is 
24  000  k.w.,  at  which  time  38  trains,  of  varying  speeds,  weighing  in  all 
9  800  tons,  are  assumed  to  be  in  motion.  The  annual  output  is  expected 
to  aggregate  121000  000  kw-hr.,  of  which  107  000  000  kw-hr.  are  for 
the  propulsion  load  and  the  remainder  for  lighting  and  other  purposes. 
These  figures  do  not  include  the  fviture  additional  requirements  for 
switching  at  various  yards,  the  movement  of  freight  trains,  and  the 
operation  of  labor-saving  devices  at  terminals. 

11  000-Volt  Transmission  Line. — The  11 000-volt  alternating  current 
from  the  power  stations  is  led  to  the  sub-stations  by  duplicate  systems 
of  insulated  copper  cables  in  ducts  within  the  populous  districts  of  the 
city;  and  by  bare  copper  cables  suspended  on  substantial  steel  poles  set 


PLATE  VIM. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN   ZONE 

OF  N.  Y.  C.  &  H.  R.  R.  R. 


PLATE  IX. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILGUS  ON 

ELECTRIFICATION   OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  &.  H.  R.  R.  R. 


PLATE  X. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILGUS  ON 

ELECTRIFICATION  OF  SUBURBAN   ZONE 

OF   N.  Y.  C.  &   H.  R.  R.  R. 


Papers.]  eLECTKIFICATION,  SUBUKBAN  ZONE,  N.  Y.  C.  &  H.  E.  R.  R.       75 

in  concrete  bases  in  the  less  densely  settled  districts.  This  arrange- 
ment was  adopted  only  after  an  exhaustive  investigation  of  line  con- 
struction throughout  the  country  had  proven  the  greater  safety  and 
reliability  of  well-built  aerial  wires,  where  the  population  is  sparse  and 
the  line  is  located  on  private  right  of  way. 

Where  the  cables  pass  from  one  type  of  construction  to  the  other, 
they  are  led  through  brick  towers  equipped  with  lightning  arresters. 

Owing  to  the  failure  of  the  city  to  grant  the  right  to  place  the 
cables  beneath  the  surface  of  neighboring  streets,  it  was  necessary  to 
locate  them  within  the  right-of-way  limits  of  the  company,  and  this 
required  many  varied  types  of  construction,  often  taxing  the  ingenuity 
of  the  engineers  to  place  the  conduit  pipes  where  they  would  be  safe 
from  injury.  A  few  of  these  conditions  are  illustrated  in  the  accom- 
panying photographs.  Altogether,  there  will  be  16  miles  of  conduit 
territory,  and  46  miles  of  pole  lines,  together  with  383  splicing 
chambers. 

Substations. — There  are  to  be  eight  sub-stations,  four  of  which 
are  now  in  operation.  Their  total  normal  rotary  capacity  will  be 
27  000  k.w. 

Each  station  contains  transformers  for  reducing  the  voltage  from 
11  000  volts  primary  to  450  volts  secondary,  and  rotary  converters  for 
changing  the  current  from  alternating  to  direct  at  660  volts.  Storage 
batteries,  "floating  on  the  line,"  are  also  provided,  to  regulate  the  sharp 
fluctuations  of  the  peculiarly  severe  short-period  demands  incident  to 
heavy  traction  service,  and  to  safeguard  the  continuity  of  traffic  should 
perchance  the  supply  of  current  be  interrupted  by  power  station  or 
distributing  failures.  This  insurance  of  reliability  of  service  has 
already  demonstrated  the  wisdom  of  its  adoption.  The  aggregate 
momentary  capacity  of  the  batteries  will  be  37  786  k.w.,  with  an  hourly 
capacity  of  12  595  k.w. 

660-Volt  Feeder  System. — The  660-volt  direct-current  system,  for 
conveying  energy  to  the  working  conductors,  consists  of  copper  cables 
protected  and  arranged  similarly  to  the  high-tension  lines  already 
described. 

Working  Conductors. — The  working  conductors  deliver  660-volt  cur- 
rent to  locomotives  and  cars.  Third-rail  is  used  at  all  points,  except 
where  intricate  switch  lay-outs  prohibit  a  continuous  conductor  near 
the  level  of  the  track.    At  such  places  overhead  conductors  are  used. 


76 


ELECTRIFICATTONT,  SVEUEBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.  [Papers. 


PLATE   XI. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  &  H.  R.  R.  R. 


Papers.]  ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.       77 

either  of  a  temporary  character  where  future  track  changes  are  con- 
templated, or  permanently  suspended  from  overhead  bridges  and  build- 
ings. The  adopted  type  of  third-rail  is  unique,  for  the  reason  that  the 
current  is  collected  from  beneath  instead  of  from  the  top.  This  permits 
the  sides  and  upper  parts  of  the  rail  to  be  sheathed  in  wood  or  other 
insulating  material  in  a  way  that  safeguards  employees  and  others  from 
accidental  contact,  and  protects  the  contact  surface  from  sleet  and 
snow  which,  with  the  usual  types  of  top-contact  rail,  so  frequently 
cause  tie-ups  of  traffic.  The  manner  of  construction  is  such  as  to 
secure  all  these  advantages,  without  encroachment  within  the  clearance 
lines  of  the  steam  equipment,  and  without  precluding  the  interchange 
of  electric  equipment  with  other  lines  already  using  the  top-contact 
type. 

At  frequent  intervals,  the  direct-current  cables  pass  through  small 
circuit-breaker  houses,  in  which  circuit-breakers  automatically  open 
and  interrupt  the  flow  of  current,  when,  because  of  accident  or  injury, 
there  is  an  improper  leak  in  the  third-rail  system  or  the  direct-current 
feeder  system.  This  safety  device,  therefore,  automatically  checks  the 
delivery  of  current  to  the  working  conductors,  when  a  continuation 
of  the  supply  might  be  disastrous.  The  circuit-breakers  are  controlled 
by  cables  connected  with  neighboring  sub-stations.  Numerous  other 
precautionary  measures  have  been  taken  for  shutting  off  power  promptly 
in  case  of  accident,  such  as,  for  instance,  continuous  indicator  wires 
for  each  of  the  four  tracks  in  the  Park  Avenue  Tunnel,  that  enable 
the  power  to  be  shut  off  immediately  on  any  desired  track. 

In  all,  there  will  be  52  miles  of  territory,  embracing  285  miles  of 
track  equipped  with  third-rail,  of  which  more  than  one-third  is  com- 
pleted and  in  use. 

Track  Bonds. — The  bonding  of  the  track  rails  for  the  return  cur- 
rent was  a  task  of  considerable  proportions,  because  of  the  intimate 
relation  of  the  work  to  traffic.  Several  ingenious  devices  were  used  in 
expediting  the  drilling  of  rails  and  placing  the  bonds.  The  concealed 
type  of  bond  was  used  as  a  protection  against  the  thefts  that  embarrass 
traffic  and  entail  pecuniary  loss,  and  to  obviate  injury  by  trackmen. 

Electrical  Equipment. — The  electrical  equipment  now  in  use  com- 
prises 35  locomotives  and  180  suburban  cars.  Of  the  cars,  125  are 
equipped  with  motors.  The  remainder,  for  the  present,  act  as  trailers, 
although  motors  will  be  added  when  the  electrical  service  is  extended 


78       ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.  [Papers. 


SPEED  AND  ACCELERATION  CURVES 
ELECTRIC  LOCOMOTIVES. 


2.0 


100 


300  300  400  500 

Total  Weight  of  Cars,  in  Tons. 

Fig.  i. 


GOO 


PLATE   XII. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  &  H.  R.  R.  R. 


Fig.  1.— Electric  Locomotive. 


Fig.  3.— Bi-polar  Gearless  Motor. 


Papers.]  ELECTRIFICATIOISr,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.       79 

the  full  distance  to  Harmon  and  North  White  Plains.  The  aggregate 
normal  rating  of  both  classes  of  equipment  is  127  000  h.p. 

Locomotives. — The  locomotive  is  a  peculiarly  efficient  and  powerful 
machine.  Although  weighing  94.5  tons,  complete,  as  compared  with 
the  171-ton  weight  of  the  heaviest  'teteam  passenger  locomotives  in  use 
by  the  company,  its  normal  rating  of  2  200  h.p.  is  practically  twice  that 
of  its  rival;  it  has  76^  tons  less  weight  to  haiil  about,  thus  effecting  a 
saving  of  45%  for  energy  in  moving  dead  tonnage;  its  concentrated 
weight  per  driving  axle,  34  250  lb.,  is  27%  less  than  that  of  the  steam 
locomotive,  without  decreasing  the  total  driver  weight  available  for 
traction;  it  is  capable  of  running  at  will  in  either  direction,  without 
the  delays  and  expense  of  going  to  the  turn-table;  it  occupies  little 
more  than  half  the  track  space  of  the  steam  locomotive — an  important 
advantage  in  terminals — and  it  is  much  more  quickly  started  and 
stopped.  These  advantages  have  been  demonstrated  strikingly  in  prac- 
tice, both  in  comparative  trials  on  the  6-mile  experimental  track  near 
Schenectady,  where  all  the  new  equipment  was  tested  exhaustively 
before  acceptance,  and  in  regular  service  in  the  New  York  zone. 

The  principal  characteristics  of  the  locomotive  are : 

Length  over  all 37  ft.     0    in. 

Rigid    wheel    base 

Total  wheel  base 

Diameter  of  drivers 

Diameter  of  truck  wheels ^ 

Total   weight 

Weight   on  four  drivers 

Weight  on  two  trucks 

Horse-power  per  ton  of  weight — normal  capacity .... 
Horse-power  per  ton  of  weight — overload  capacity.  . 

Number  of  motors 

Normal  capacity  of  each  motor 550  h.p. 

Normal  capacity  of  each  locomotive 2  200 

Over-load  capacity  of  each  locomotive 3  300 

Type  of  motors Gearless,  bi-polar. 

Type  of  control Sprague-General  Electric  multiple-unit. 

Type  of  heaters  for  train  supply Westinghouse    oil-fired. 

Air  brakes Westinghouse  graduated-release. 


37  ft. 

0 

13    " 

0 

27    " 

0 

44 

36^ 

94^  tons. 

68A 

a 

26 

(I 

23 

35 

4 

62 

ft. 

0 

m. 

50 

iC 

0 

38 

a 

6 

7 

u 

0 

6 

u 

0 

36 

33 

80        ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.   [Papers. 

Cars. — The  suburban  cars  are  constructed  of  steel  and  other  non- 
inflammable  material,  and,  while  simple  in  design,  have  all  the  features 
conducive  to  the  safety  and  comfort  of  the  public.  Their  leading  char- 
acteristics are  as  follows : 

Length,  over  all 62  ft. 

Length  of  car  body 

Distance  between  truck  centers 

Distance  between  axles  of  motor  trucks 

Distance  between  axles  of  trailer  trucks 

Diameter  of  wheels — motor  trucks 

Diameter  of  wheels — trailer  trucks 

Number  of  motors  on  each  motor  truck 2 

Normal  capacity  of  each  motor 200  h.p. 

Normal  capacity  of  motor  car 400    " 

Total  weight  of  motor  car 53     tons. 

Total  weight  of  trailer  car 44^      " 

Total  weight  of  car  body 33J      " 

Weight  per  motor  car,  due  to  electrical  equipment.  ...         SJ      " 
Horse-power   (normal  capacity)   per  ton  of  weight  of 

electrical   equipment 47 

Seating  capacity 64 

Heating  system Both  steam  and  electric. 

Lighting  system Both  electric  and  Pintsch  gas. 

Cooling  system  for  summer  season Two  14-in.  electric  fans. 

Type  of  control Sprague-General  Electric  multiple-unit. 

Acceleration,  in  miles  per  hour  per  second 1.2 

Comparative  Train  Weights. — The  comparative  weights  of  steam 
and  electric  trains  in  the  two  classes  of  service,  through  and  suburban, 
are  interesting,  as  illustrative  of  the  saving  in  consumption  of  energy, 
and  therefore  in  cost  of  operation,  that  accompanies  the  lower  electric 
train  weights;  and,  also,  as  justifying  the  adoption  of  the  multiple-unit 
instead  of  locomotive  practice  for  suburban  operation. 

Through  Service. 

Steam.  Electric. 

Tous.  Tons. 

Pacific  type  locomotive .  .     171.0  Electric  locomotive 94.5 

8  Pullman   cars 400.0  8  Pullman  cars 400.0 


Total 571.0  Totiil 494.5 

Saving  in  favor  of  electric  traction  =  7Gi  tons  =  13  per  cent. 


PLATE  XIII. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF   N.  Y.  C.  &   H.  R.  R.  R. 


ELEcriue  LoetumTtye  S.3CjMt  TfuUft-TOrM.  lirT/94./S  Toms. 


/e*-e/ //c^c^  ^r^d rt?presenf  ai-srsg^  coni/jfions  o^ fregfher 
Corrocfions  muof  be  maefe  for gr^t/e  3ncy cur^aft/ns  of  /rcKff 


Fia.  1.— Train  Braking  Chart— Distance. 


SL£:cr/f/cLocoMOr/y£-S  6  Cj^r  TffAm-ToTAL  ft^T.  2^338  Tons. 


Corrections  musf  be/773c/^  forQrscfe3n<fcari'3fLfre  offriScA 
tr/7ef7  3pp//ecf  fo  snij  ^/i^e/7  /ooaf/on.  Carres  ^ns  p/offec^  from  ^n 
syeraffe  o/'t^  /ri/mher  of  runs. 


Fig.  2.— Train  Brakfng  Chart— Time. 


Papers.]  ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.       81 

Suburban    Service. 

Average  Number  of  Cars. 

Electric   Locomotive.  Multiple-Unit  Cars. 

Tons.  Tons. 

Locomotive    94.5 

4i  steel  trailer  cars 200.0  4^  motor  cars 238.5 

Total 294.5  Total 238.5 

Saving  in  favor  of  multiple-unit  practice  =  56  tons  =  19  per  cent. 

Shops  and  Inspection  Sheds. — The  maintenance  of  electrical  equip- 
ment in  a  high  degree  of  efficiency  requires  suitable  repair  shops  and 
inspection  sheds,  located  where  the  dead  mileage  will  be  reduced  to  a 
minimum.  At  both  Harmon  and  North  White  Plains  permanent  in- 
spection sheds  have  been  built,  and  at  the  former  point  ample  modern 
shop  facilities  are  provided.  As  the  equipment  on  both  divisions  is 
pooled,  any  car  or  locomotive  needing  repairs  can  be  sent  while  in 
regular  service  to  either  place,  without  the  expense  and  loss  of  time 
incident  to  special  dead  movements. 

Interchange  Terminals. — At  North  White  Plains,  the  existing  steam 
engine-house  plant  is  to  be  enlarged  when  the  extension  of  electric 
operation  requires  added  facilities  for  the  interchange  of  power.  At 
Harmon,  space  has  been  provided  for  ample  facilities  for  the  same 
purpose.  Owing  to  the  present  curtailment  of  electric  operation  because 
of  the  backwardness  of  the  State  in  acting  on  the  abolition  of  grade 
crossings  north  of  the  limits  of  the  City  of  New  York,  temporary 
terminals  have  been  constructed,  at  High  Bridge  on  the  Hudson 
Division  and  at  Wakefield  on  the  Harlem  Division,  with  convenient 
yard  arrangements  and  structures  for  the  care  and  exchange  of  power. 

Operating  Organization  of  Electrical  Department. — The  success  of 
a  new  plant  of  such  magnitude,  especially  when  a  change  from  old  to 
new  conditions  must  be  effected  without  embarrassing  an  enormous 
passenger  traffic,  depends  very  largely  on  the  organization  and  per- 
sonnel of  the  electrical  operating  force.  It  was  recognized,  at  an  early 
stage  of  the  work,  that  the  operation  and  maintenance  of  the  entire 
installation  required  to  deliver  current  to  equipment,  as  well  as  the 
maintenance  of  locomotives  and  cars,  should  be  under  the  supervision 
of  those  responsible  for  their  construction,  leaving  to  the  regular  steam 


82       ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.  [Papers. 

organization  the  operation  of  trains  with  electric  current  and  equip- 
ment thus  furnished. 

As  the  work  on  the  power  stations,  distributing  system,  and  equip- 
ment progressed,  competent  men  were  gradually  employed  for  inspec- 
tion and  testing  purposes,  so  that,  when  all  was  ready  for  regular 
■operation,  there  was  in  existence  a  skilled,  energetic  corps  of  veterans, 
■equal  to  any  emergency,  and  imbued  with  a  spirit  that  meant  success. 

Telephone  System.- — A  word  should  here  be  spoken  of  the  independ- 
ent telephone  system  which  has  been  constructed  for  the  purpose  of 
bringing  all  parts  of  the  electric  zone  in  close  touch  with  each  other 
and  with  the  load  and  train  dispatchers. 

Other  Improvements. — While  the  principal  purpose  of  this  paper 
is  to  give  an  outline  of  the  elements  of  the  electrification  of  the  New 
York  Central  suburban  zone,  it  would  be  incomplete  without  at  least  a 
passing  mention  of  the  other  important  improvements  undertaken  in 
conjunction  with  the  change  of  motive  power. 

Grand  Central  Terminal. — Within  the  territory  bounded  by  Forty- 
second  Street,  Fifty-seventh  Street,  Madison  Avenue  and  Lexington 
Avenue,  the  old  Grand  Central  Terminal  occupied  a  parcel  of  irregular 
shape,  with  an  area  of  about  23  acres.  The  four  main  tracks  from  the 
north  descend  on  grades  of  from  26  to  53  ft.  per  mile  to  the  south  end 
cf  the  Park  Avenue  Tunnel  at  Fifty-sixth  Street;  thence  they  ascend 
at  the  rate  of  62  ft.  per  mile  in  an  open  cut  in  the  middle  of  Park 
Avenue  to  Fiftieth  Street;  thence  spreading  out  into  the  yard,  on  a 
slight  descent  to  Forty-fifth  Street;  and  thence  on  a  gentle  declivity 
to  the  terminal  in  the  train-shed  near  Forty-third  Street.  The  vital 
defect  of  this  arrangement  was  the  absence  of  switching  tracks  for 
drilling  the  yard,  north  of  Fiftieth  Street,  which  necessitated  the  use 
of  two  of  the  main  tracks  for  that  purpose.  Consequently,  the  entrance 
to  the  terminal  really  consisted  of  but  two  main  tracks  for  the  accom- 
modation of  the  trafiic  pouring  to  and  from  a  four-track  line.  To 
increase  the  congestion,  one  of  these  tracks,  assigned  to  drilling  ser- 
vice, had  also  to  be  used  for  the  storage  of  steam  locomotives  at  rush 
hours  of  the  day. 

By  the  use  of  electricity,  it  became  possible  to  depress  the  roadbed 
south  of  the  low  point  at  Fifty-sixth  Street,  so  as  to  pass  beneath  the 
surface  of  Park  Avenue  on  either  side  of  the  railroad,  and  thus  permit 
the  utilization  of  the  full  width  of  the  avenue,  140  ft.,  without  affecting 


PLATE  XIV. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILGUS  ON 

ELECTRIFICATION  OF  SUBURBAN   ZONE 

OF  N    Y.  C.  &   H.  R.  R.  R. 


Fig.  3.— High  Bridge  Temporary  Terminal. 


Fi(i.  3.— Wakefield  Temporary  Terminal. 


Papers.]  eLECTIUFICATION,  SUBUU15AN  ZONL,  N.  Y.  C.  &  11.  U.  R.  R.       83 

its  use  by  the  public.  This  gave  space  for  ten  instead  of  four  tracks 
from  Fifty-sixth  to  Fiftieth  Streets,  of  which  four  are  for  a  legitimate 
main-line  entrance  to  the  enlarged  upper  yard,  two  are  for  drilling  the 
yard,  and  two  on  each  side,  or  four  in  nil,  ;ire  for  ingress  and  egress 
of  the  lower-level  suburban  station.  The  upper  level  for  through  trains 
will  have  stub  tracks,  while  the  lower  level  will  have  a  double-track 
loop  at  the  south  end  near  Forty-third  Street. 

The  depression  also  admits  of  the  extension  of  Park  Avenue,  for 
its  full  width,  south  from  Fiftieth  to  Forty-fifth  Streets  over  the  tracks 
of  the  yard,  and  the  connection  by  east  and  west  viaducts  of  the  ends 
of  streets  from  Forly-Hflh  to  Fifty-sixth  Strcx^ts.  inclusive,  now  sepa- 
rated by  the  lonninnl. 

To  the  21?  acres  in  tlio  old  (onniiial  lins  boon  nddod  by  luirchase  17 
acres,  nuiking  a  total  area  of  40  acres.  With  the  21  acres  obtained 
by  excavating  for  the  suburban  stiUion,  there  will  be  a  total  area  in  the 
new  terminal,  when  com]ileted,  of  more  than  6i  acres.  This  is  equal 
to  an  increase  over  the  present  space  of  178  per  cent. 

These  radical  changes  make  necessary  the  tearing  down  of  the  old 
station  and  train-shed,  originally  built  in  1871  and  enlarged  in  1898 
and  1900;  and  the  substituticm  of  a  nuicli  larg(M-  and  handsomer  struc- 
ture, suited  to  the  new  motive  power  and  more  adequate  for  the  proper 
handling  of  a  rapidly  increasing  tratlic. 

It  should  here  be  added  that  electricity  brings  with  it  an  unexpected 
boom  in  the  permissible  use  of  overhead  spaces  termed  "air  rights," 
that  is  denied  with  steam  traction.  A  vast  area  in  the  heart  of  the 
greatest  city  on  the  continent  is  thus  reclaimed  for  use  as  desired  for 
various  revenue-producing  purposes.  In  time,  lliis  feature  will  add 
very  largely  to  the  company's  assets. 

An  idea  of  the  difficulties  of  construction,  due  to  the  nature  of  the 
underlying  material — solid  rock — and  the  necessity  of  subordinating 
all  efforts  to  the  safe  and  uninterrupted  movement  of  an  exacting  and 
constantly  increasing  train  service,  is  illustrated  in  the  accompanying 
photographs. 

Tlie  magnitude  of  the  now  lerniinal,  wliich  has  thus  to  be  built 
while  trains  and  passengers  i)onr  in  and  out,  is  evident  from  the  quan- 
tities of  material  involved.  A  Iter  Iteiug  loaded  on  cars,  3  000  000  cu.  yd. 
of  rock  and  earth  are  dispatclied,  at  times  when  the  passenger  service 
will    j)G™^it,    to    the    Hudson    Division,    for    building   additional    main 


84       ELECTEIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.  [Papers. 

tracks.  In  the  construction  of  retaining  walls,  suburban  stations,  via- 
ducts, subways,  and  tunnels,  100  000  tons  of  steel  and  260  000  cu.  yd. 
of  concrete  are  used,  in  addition  to  numerous  other  materials. 

The  final  result  will  be  an  electrically  operated  station  and  yard 
with  quadruple  the  capacity  of  the  old  one,  and  with  many  appurte- 
nances for  usefulness  and  profit,  that  are  additional  to  the  parent 
purpose  of  a  railroad  terminal. 

Bronx  Improvement. — At  Mott  Haven  Junction,  in  the  Borough  of 
the  Bronx,  5.3  miles  from  the  Grand  Central  Terminal,  two  four-track 
lines,  making  eight  tracks  in  all,  merge  into  the  single  four-track  stem 
that  leads  to  the  terminal.  With  increased  frequency  of  train  service,^ 
the  grade  intersections  at  such  an  important  junction  are  inadmissible 
on  tlie  grounds  of  safety  and  non-delay  to  traffic.  Therefore,  plans 
have  been  adopted  and  work  commenced  on  the  raising  and  lowering 
of  tracks  by  means  of  viaducts  and  tunnels,  so  as  to  effect  trailing 
junctions  free  from  grade  crossings. 

The  points  of  junction  are  to  be  moved  south  about  |  mile,  to  the 
vicinity  of  the  Harlem  River;  and,  near  the  present  connection,  on 
One  Hundred  and  Forty-ninth  Street,  a  new  large  overhead  station  is 
to  be  built,  with  eight  main  tracks.  This  will  permit  the  abandonment 
of  the  old  station  at  One  Hundred  and  Thirty-eighth  Street,  and 
remove  another  of  the  causes  for  congestion  on  the  four-track  entrance 
to  the  Grand  Central  Terminal.  Moreover,  this  new  station  will  serve 
the  rapidly  growing  population  in  the  Bronx,  which  is  fast  approaching 
the  half-million  mark. 

Elimination  of  Grade  Crossings. — At  the  time  of  the  decision  to 
proceed  with  electric  zone  improvements,  there  were,  within  that  terri- 
tory, forty-four  street  and  highway  grade  crossings,  the  abolition  of 
which  was  deemed  precedent  to  the  commencement  of  electric  opera- 
tion. Of  these,  one-half  were  located  within  the  city  limits  of  New 
York,  and  these,  by  agreement  with  the  City,  have  since  been  carried 
over  the  tracks.  None  of  the  remainder  has  yet  been  completed,  owing 
to  the  delay  of  the  State  authorities  to  make  effective  the  provisions  of 
the  statute  governing  grade  crossings,  and  also  owing  to  difficulties  in 
acquiring  the  necessary  additional  right  of  way.  However,  due  to  the 
energetic  action  of  the  new  Public  Service  Commission,  decisions  on 
many  of  the  crossings  have  been  reached;  and  the  remainder  are 
expected   soon.      The   majority    of   these    eliminations    require    either 


PLATE  XV. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION   OF  SUBURBAN  ZONE 

OF   N.  Y.  C.  &   H.  R.  R.  R. 


Papers.]   ELHCTlil  I'lCATION,  SUBUHBAN  ZONK,  N.  Y.  C.  &  II.  U.  K.  It.        85 

changes  of  the  line  of  the  railroad  for  considerable  distances,  as  for 
instance  at  Mount  Vernon  and  White  Plains,  or  the  lifting  of  the 
grade  of  the  tracks  so  that  the  streets  may  pass  under,  as  at  Yonkers 
and  Tarrytown. 

Local  Improvements. — The  elimination  of  grade  crossings  north  of 
New  York  City,  and  the  growth  of  business,  make  obligatory  many 
extensive  and  costly  local  improvements,  with  new  passenger  and 
freight  stations  and  yards.  The  more  important  ones  are  at  Yonkers, 
Hastings,  Tarrytown,  Ossining,  and  Harmon,  on  the  Hudson  Division; 
and  at  Mount  Vernon,  Bronxville,  Tuckahoe,  and  White  Plains,  on 
the  Harlem  Division.  Features  of  the  design  of  these  new  stations 
are  the  avoidance,  by  means  of  subways  and  overhead  bridges,  of  all 
grade  crossings  of  tracks  by  passengers;  and  the  placing  of  the  tops 
of  the  local  platforms  on  a  level  with  the  car  floor, 

Four-Tracking  and  Loops. — The  anticipated  increase  in  frequency 
of  train  service  with  electric  traction,  and  the  urgent  necessity  of 
removing  causes  of  congestion  in  this  important  entrance  to  New  York 
City,  make  mandatory  the  construction  of  additional  main  tracks,  so 
that  there  will  be  separate  tracks  in  each  direction  for  high-  and  slow- 
speed  service;  and,  where  possible,  additional  tracks  for  the  exclusive 
movement  of  freight. 

In  line  with  this  policy,  new  main  tracks  are  under  construction 
within  the  suburban  zone,  in  conjunction  with  the  elimination  of  grade 
crossings  and  improvement  of  local  facilities.  The  four  tracks  on  the 
Harlem  Division  are  being  extended  from  Woodlawn  Junction  to  North 
White  Plains,  with  long  middle  sidings  at  frequent  intervals,  for  the 
passage  of  passenger  trains  around  freights.  The  double  and  triple 
main  tracks  on  the  Hudson  Division,  as  far  out  as  Harmon,  are  being 
increased  to  four,  and,  at  some  places,  as  for  instance  between  Spuyten 
Duyvil  and  Yonkers,  two  additional  tracks  have  been  provided  for  the 
exclusive  use  of  freight  trains.  As  on  the  Harlem  Division,  middle 
tracks  are  being  built,  where  needed,  for  keeping  freight  trains  out  of 
the  way  of  the  passenger  service. 

At  Harmon  and  North  White  Plains,  loops  are  to  be  built,  for  the 
turning  of  suburban  trains  without  crossing  the  express  traffic  at  grade. 
It  will  be  noted  that,  with  loops  at  all  three  termini  and  the  freedom 
from  grade  crossings  at  Mott  Haven  Junction,  opportunity  is  given  for 
a  constant  flow  of  traffic  with  an  absence  of  the  usual  obstructions  that 
cause  congestion. 


8G       ELECTRIFICATION,  SUBURBAN  ZONE,  N.  T.  C.  &  H.  R.  R.  R.  [Papers. 

Increased  Capacity  of  Entrance  to  Grand  Central  Terminal. — From 
the  fact  that  two  four-track  lines  feed  into  a  single  four-track  stem 
from  Mott  Haven  Junction  to  the  terminal,  the  question  naturally 
arises  as  to  what  solution  the  future  holds  for  this  restriction  on 
growth  of  traffic.  The  present  plans  of  the  terminal  provide  for  a 
future  four-track  cross-town  tunnel  connection  with  the  West  Side 
line  of  the  Company,  over  which  the  Hudson  Division  can  then  enter 
the  terminal  without  burdening  the  Harlem  Division  tracks.  This, 
when  built,  will  afford  to  the  terminal  an  eight-track  entrance  con- 
nected with  both  train  levels. 

Improvements  in  Alignment  and  Grades. — In  conjunction  with 
these  radical  "changes  in  the  physical  condition  of  the  property,  it  has 
been  considered  wise  to  make  at  the  same  time  other  desirable  changes 
that  could  not  be  accomplished  later  without  undue  extra  cost.  At 
many  places,  on  both  divisions,  alignment  and  grades  have  had  careful 
study,  and  alterations  have  been  approved  which  will  result  in  material 
saving  in  rise  and  fall,  and  in  curvature.  Many  have  been  completed, 
and  others  have  been  deferred,  awaiting  the  acquisition  of  right  of  way 
and  the  settlement  of  legal  questions.  Among  those  still  in  embryo  is 
the  improvement  between  Croton  and  Peekskill,  more  than  8  miles  in 
length,  which  when  completed  Avill  admit  of  a  still  further  extension  of 
electric  operation. 

The  advantages  to  be  gained  by  the  principal  changes  of  alignment 
are  as  follows : 


Marble    Hill    cut-off,    including    Spuyten 

Duyvil   Tunnel  cut-off 

Croton    to    Peekskill 

Spuyten  Duyvil  to  Mt.  St.  Vincent 

Irvington    cut-off 


Totals 8  341    ''  559° 

Signals  and  Interlocking. — Under  the  old  order  of  affairs,  traffic  on 
the  Hudson  Division  from  the  north  ran  right-handed  to  Spuyten 
Duyvil,  where  it  was  transposed  to  left-handed  operation  so  as  to 
harmonize  with  the  left-handed  practice  on  the  Harlem  Division.  The 
design  of  the  new  Grand  Central  Termiuiil  and  a  possible  future  con- 


Saving  in 

Saving  in 

Distance. 

Curvature, 

3  944  ft. 

137° 

4  338    " 

333° 

9    " 

24° 

50    " 

65° 

PLATE   XVI. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN   ZONE 

OF  N.  Y.  C.  &   H.  R.  R.  R. 


1     '1 

! 

►'k      ■   %■  9=*^ 

j; 

L-.     1 

b  :  Vr 

J^^^ 

-'  MiSk  \ 

I- 

iSS^iRASl 

}- 

V 

^^Kxw^fjiall 

Tl 

'i|H|'.  1 

4- 

i,     /^^SKP 

33 

,  ^Hfe:  ' 

'^"Si 

1 ,  w^  ■ 

K 

1  s     -^^ 

-  *  ^\^l^ 

^■^ 

1  itflK 

H 

i  1  V 

o 

k"^ 

a.  1  IB                 „^^mM 

w 

"  ■" ' ' — ^ 

* 

'tfr  -«Hir*>i' ' 

>■ 

0 

o 

""^j 

'  -9 

H 

-^ ,   ,-£y» 

s-J 

"    M.-. 

z 

jxmmtik^i  ,  1 

H 

^^^^^v  "'Sw    .  vftV 

mKKHmd4' 

t- 

si^^^^k 

H 

^^^^  1^1 ** 

H 

^^^^^^^B^rF? 

1> 

./jMy 

M^^/^yyimm 

5; 

|Ufi^V 

mKk/////WM 

MH^^^H 

<i  vQjfi^RluSik. ' 

ffl 

^iSS^wvnu. 

M 

wSfmSw^^k^    ' 

O 

„..'.^        1* 

^ 

l';i|)cis.]    |.;ij.;("i'|;|  |'|(!A'I'I()N,  SIHUIL'l'.A  N   /ONI'l,  N.   Y.  (!.  &  11.  If.   U.  It.        87 

nection  with  the  cil.y  siibwny  h.vsIciii,  as  well  as  (he  dcsirMhilily  of 
avoiding  tiic  gradt^  crossing  at  Spuylcn  Dnyvil,  led  l,o  tho  decision  to 
tnako  llio  riglit-lianded  system  of  operation  uniform  throughout  the 
Hiihiirlian  zone.  In  considering  Ihr  eifect  of  this  reversal  of  traffic 
on  the  existing  signals  and  interlocking  plants,  it  was  also  i-ealized 
that  the  cuntrollcd-manual  systc^m  in  use  on  a  largo  portion  of  the 
territory  was  insufTieicuitly  elastic  for  tho  quick  handling  of  a  frequent 
electric  train  scrvieo  on  four  oi-  nioi'e  tracks.  Aeeompanying  these 
traffic  reasons  for  radical  t^haiigcs  in  tho  old  signals  and  interlockings 
was  the  ecpially  important  fact  that  the  use  of  track  rails  for  return 
pro|)ulsi()n  current,  to  tlu!  i)ovv(!r  stations  completely  dc'ranged  the 
signal  circuits.  Thcji,  too,  the  many  additions  and  changes  to  tracks 
made  imperative  the  abandonment  of  tho  larger  ])art  oi,  the  old  plants. 

All  these  causes  led  to  the  adoption  of  new  electric  automatic  sig- 
nals and  electric  interlocking  plants  for  the  entire  zone,  the  predomi- 
nant feature  of  which  is  the  reactance  bond,  which  p(!rmits  the  free 
j)assage  of  propulsion  ciirreiit,  through  the  track  rails,  but,  where  de- 
sired, stops  the  passage  of  the  altcM-nafing  signal  cnri'cait  ci remit. 

Fences. — One  of  the  worst  evils  with  which  American  railroads 
have  to  contend  is  trespass.  Kight  of  way  and  tracks  are  considered 
public  highways,  and  tho  petty  courts  refuse  or  neglect  to  impose 
adequate  punishment  on  those  who  thus  risk  their  lives  in  dangerous 
places.  This  freedom  of  use  of  the  railroad's  property  also  leads  to 
thieving  which,  in  the  aggregate,  causes  large  losses  to  the  company. 
'riw.  change  to  electric  traction  by  no  means  minimizes  these  evils. 
More  frequent  trains,  and  the  presence  of  electricity,  increase  the  risks, 
while  copper  cables  and  bonds  attract  the  thief.  To  guard  against  Ihese 
incrcas(!d  dangers,  tb(!  entire;  ((l(H;tric  zone  is  to  be  enclosed  with  num- 
and  boy-proof  fences.  The  portion  within  the  settled  districts  consists 
of  iron  pickets  and  concrete  posts  of  a  pleasing  design. 

Chronology.— -YoWowmg  the  placing  of  orders  in  tli(!  fall  of  V.)OS, 
work  was  pushed  energetically  on  all  items  of  construction  required 
for  tho  operation  of  the  initial  electric  zone  south  of  Wakefield  and 
High  Bridge.  It  should  be  borne  in  mind  that  the  larger  part  of  the 
work  had  to  be  p(!rfortn(ul  on  or  about  tracks  congested  with  traffic, 
which  entailed  danger  to  employees,  dcjiay  to  numy  parts  of  the  work, 
and  ex])ense.  It  is  a  pleasure  to  record,  however,  that  not  an  accident 
occurred  to  regular  train  service,  nor,  with  a  lew  minor  (!Xce|)l  ions,  any 


88        ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.  [Papers. 

delay  to  traffic,  due  to  construction.    Such  accidents  and  delays  as  did 
occur  were  from  other  causes. 

The  following  dates  mark  the  progress  of  the  electrical  features : 

Initial  informal  test  of  first  electric  loco- 
motive    October       27th,  1904. 

First  formal  test  of  Electric  locomotive.  . .  .  November  12th,  1904. 

Port  Morris  Power  Station : 

Commencement    May  15th,  1904:. 

First   current July  1st,  1906. 

Transmission  Lines: 

Commencement   February    17th,  1905. 

Ready  for  service September  30th,  1906. 

Sub-stations : 

Commencement    July  6th,  1905. 

Ready  for   service September  30th,  1906. 

Working  Conductors : 

Commencement   January         2d,  1906. 

Ready  for  service December   11th,  1906. 

Electrical  Equipment : 

First  operated  in  New  York   City...  July  20th,  1906. 
First   train    into   Grand    (^entral    Ter- 
minal      September  30th,  1906. 

Electrical  Operation : 

First  schedule  multiple-unit  train.  ..  .  December    11th,  1906. 

First  schedule   electric  locomotive.  ...  February    13th,  1907. 

First  regular  shop  train April  14th,  1907. 

Completion  of  change  of  motive  power: 

Schedule   trains July  1st,  1907. 

Reversal  of  traffic August       25th,  1907. 

Because  of  the  burdensome  conditions  of  traffic,  and  complicated 
changes  in  the  signal  and  interlocking  systems,  about  6  months  were 
thus  consumed  in  making  the  change  of  motive  power  complete,  after 
the  first  schedule  train  was  operated. 


PLATE  XVII. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  &  H.  R.  R.  R. 


Fig.  1.— Before. 


Fig.  3.— After. 

Contrast   Between  the  S.vioke  Conditions,  as  They  Existed  at  the  Grand  Central 

Terminal  in  1906.  and  the  Absence  of  S.moke  in  the  New  Terminal. 

Due  to  the  Use  of  Electricity. 


Papers.]  ELECTRIFICATION,  SUBURBAN  ZONE,  N,  Y.  C.  &  H.  R.  R.  R.       89 

Initial  Zone  Operation. — As  previously  stated,  the  company  was 
forced  to  confine  temporarily  the  change  of  motive  power  to  the  opera- 
tion of  the  suburban  zone  terminating  at  High  Bridge,  Y  miles  out; 
and  at  Wakefield,  13  miles  from  the  terminal.  This  postpones  for 
two  or  three  years  the  extension  of  electrical  service  to  the  northerly 
termini  of  the  suburban  zone.  In  the  meantime,  the  power  on  through 
trains  is  changed  at  the  temporary  termini.  At  the  same  points,  mul- 
tiple-unit trains  north-bound  have  steam  locomotives  attached  and 
thence  proceed  as  non-electric  trains;  and  south-bound  the  steam 
locomotives  are  detached  and  the  trains  continue  by  electricity  without 
locomotives.  The  average  time  required  for  making  the  changes, 
including  that  lost  in  slowing  down  and  regaining  speed  is  as  follows: 

Through  trains  with  locomotives 4 J  min. 

Multiple-unit  trains,  north-bound 3       " 

Multiple-unit  trains,  south-bound 2^     " 

On  the  Hudson  Division  this  delay  has  been  largely  compensated 
by  shortening  the  line  at  Marble  Hill  and  the  elimination  of  grade 
track  crossing  at  Spuyten  Duyvil. 

Eesults. 

Expectations  from  Electrification. — Now  that  the  change  of  motive 
power  in  the  initial  electric  zone  has  been  completed  for  sufficient  time 
to  gain  at  least  a  preliminary  idea  of  the  results,  the  question  naturally 
arises,  with  what  success  has  the  change  met  expectations  ? 

It  has  already  been  explained  that  the  principal  reasons  for  under- 
taking the  work  were  twofold: 

(1). — Demand  of  the  public  for  the  abolition  of  the  nuisances 
incident  to  the  use  of  steam  locomotives  south  of  the 
Harlem  River;  and 

(2). — Need  for  increased  capacity  of  the  terminal,  by  the  elimi- 
nation of  a  large  proportion  of  the  switching  movements 
required  with  steam  locomotive  practice;  and  relief  to 
the  main  line  entrance  to  the  terminal  by  reducing  its 
use  for  haulage  of  dead  locomotives  and  cars  to  Mott 
Haven. 


90       ELECTRIFICATION,  SUBUllBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.  [Papers. 

As  secondary  considerations  there  were: 

(3). — The  possibility  of  sufficient  economy  in  operation  at  least  to 
offset  largely  the  additional  fixed  charges  on  the  cost  of 
the  electrical  installation;,  and 

(4). — Opportunities  for  an  ultimate  large  increase  in  traffic  and 
corresponding  growth  of  revenue  to  justify  the  expendi- 
ture for  all  improvements  within  the  suburban  zone. 

What  do  the  observations  made  thus  far  disclose? 

The  first  two  expectations  have  been  completely  realized. 

Park  Avenue  Tunnel. — The  atmospheric  conditions  in  the  Park 
Avenue  Tunnel  show  marked  improvement,  even  with  the  presence  of 
the  remaining  New  Haven  Company's  steam  service. 

Increased  Terminal  Capacity. — The  effect  on  the  operating  effici- 
ency of  the  terminal  has  been  very  gratifying,  the  increased  capacity 
being  estimated  at  one-third.  There  has  also  been  a  large  reduction  in 
the  number  of  shop  or  "dead"  trains  to  and  from  Mott  Haven. 

Reduced  Cost  of  Opei-ation. — The  results,  as  regards  the  third  ex- 
pectation, have  been  most  surprising.  The  operation,  for  a  considerable 
period,  of  steam  and  electric  equipment  side  by  side  has  afforded  an 
unexampled  opportunity  for  a  true  comparison  of  costs  of  operation. 
Until  now,  data  on  this  subject  have  been  based  on  theory,  ignoring 
many  of  the  indeterminate  features  of  actual  operation  that  have  such 
a  weighty  effect  on  costs.  For  instance,  among  the  variables  entering 
into  an  analysis  of  this  character  are : 

(a). — Cost  and  quantity  of  coal  and  water  at  the  power  station, 
and  on  the  steam  locomotive  tender; 

(b). — Relation  of  ton-mileage  of  the  motive  power  to  total  ton- 
mileage,  including  motive  power  and  cars; 

(c). — Frequency  and  volume  of  traffic; 

(d). — Mechanical  and  electrical  desigTi  of  motive  power  as  affect- 
ing repairs,  and  hours  available  for  active  service; 

(e). — Fixed  charges,  depreciation,  and  maintenance  on  all  items 
of  both  kinds  of  service,  that  have  a  bearing  on  compara- 
tive results,  including  land,  structures,  and  equipment. 

In  other  words,  to  obtain  a  true  comparison,  observations  must  be 
made  under  like  conditions  in  a  known  service. 


PLATE   XVIII. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILGUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.Y.  C.  &  H.  R.  R.  R. 


^^rrrs^^j 


j:: 


■«ffrv„v,-.  r  iirpajra     t  ^•f'^^^^r'-  X 


Fui.  1.     I'liMi'dSKL)  liitiiNX  Station. 


<A  -Kc-  -..^ 


%' 


Fig.  2.— Orthogkaphic  View  of  Bronx  Improvement. 


Fig.  3.     Tvi'KAL   Gradk   CiicissrNc    Ki.imination.   with    (  »vi:hhioaii   m 
AND  TRACK.S  Beneath  (University  Heights). 


Papers.]  ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  II.  R.  R.  R.       91 

With  this  object  in  view,  a  typical  steam  switching  locomotive, 
engaged  in  terminal  service,  and  a  steam  passenger  locomotive,  assigned 
to  road  service,  were  each  selected  for  observation  in  the  same  class  of 
traffic  with  electric  locomotives.    The  terminal  service  embraced  switch- 

STEAM  LOCOIVIOTIVES  USED  IN  COMPARATIVE  TESTS 


Weight  on  drivers,  working  order 156  000  lb. 

Weigbt  on  truck,  working  order 16  500  lb. 

Wciglil;  total  of  eujiue _191  500  lb. 

Weighlof  tender,  loaded 118000  lb. 

STEAM   LOCOMOTIVE  USED  IN 

ROAD  TESTS. 

CLASS-F-2-d. 

(No. 1978.) 


iniX-O^ 


-16  6Ji- ! 


:>s  w}. 


Weight  on  drivers,  working  order 152  500  lb. 

Weight,  total  of  engine 152  5001b. 

Weight  of  lender,  lo.lded 

4DO0-gal.  t.ank _..S9  5001b. 

5100-gal.  t.inU 915001b. 

STEAM   LOCOMOTIVE  USED  IN 
SWITCHING  AND  HAULING  TESTS. 
CLASS  B-10. 
^    Fig.  3. 


ing  at  the  Grand  Central  yard,  and  hauling  dead  cars  to  and  from 
Mott  Haven  storage  yard,  a  distance  of  6  miles.  The  road  service 
comprised  the  hauling  of  schedule  trains  by  the  electric  locomotive 
between  the  Grand  Central  Terminal  and  Wakefield,  12^  miles;  and 


92       ELECTRIFICATION",  SUBURBAN  ZONE,  N,  Y.  C.  &  H.  R.  R.  R.  [Papers. 

the  same  trains  by  steam  between  Wakefield  and  North  White  Plains, 
11^  miles. 

Observers  constantly  rode  the  locomotives  for  the  period  of  the 
tests,  namely,  September  12th  to  27th,  1907,  in  terminal  service,  and 
October  4th  to  18th,  1907,  in  road  service.  Cyclometers  and  watt- 
meters registered  actual  distances,  speeds,  and  current  consumption. 
Record  was  also  kept  of  the  number  of  cars  switched  and  hauled,  and 
the  proportion  of  time  each  day  engaged  in  actual  service,  awaiting 
duty,  and  laid  up  for  inspection  and  repairs. 

The  coal  used  contained  14  000  B.  t.  u.  per  lb.,  and  the  cost,  per 
ton  of  2  240  lb.,  was : 

Steam  locomotive  in  terminal  service  (anthracite) ....     $5.00  per  ton. 

Steam  locomotive  in  road  service  (bituminous) 3.50     "       " 

Port  Morris  power  station  (bituminous) 3.05     "       " 

Water,  per  1  000  gal.,  cost  as  follows : 

Terminal  service  and  at  power  station 13 J  cents. 

Road  service 5        " 

The  cost  of  electric  current,  when  the  power  station  designed  load 
is  attained,  is  taken  at  2.6  cents  per  kw-hr.,  delivered  at  the  contact 
shoes  of  the  equipment,  and  includes  all  operating  and  maintenance 
costs,  interest  on  the  electrical  investment  required  to  produce  and 
deliver  the  current,  depreciation,  taxes,  insurance,  and  transmission 
losses.     The  details  of  this  cost  are : 


Items. 

Operating  costs. 

Fixed  charges. 

Total. 

Power  station 

$0.58 
0.19 

0.32 

$0.44 
0.15 

0.92 

$1.08 

Transmission  losses 

0.34 

Distributing  system  and  .-iib-sta 

1  24 

Totals 

$1.09 

$1.51 

$2.60 

Locomotive  wages   are  practically  identical  for  each  class  of  ser- 


Table  1  shows  the  details  of  locomotive  repairs,  maintenance,  and 
fixed  charges  for  each  class  of  service,  from  which  it  will  be  noted 
that,  although  the  fixed  charges  and  depreciation  of  the  electric  loco- 


PLATE  XIX. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.Y.  C.  &  H.  R.  R.  R. 


PLATE   XX. 

PAPERS,  AM.  SOC.  C.  E. 

FEBRUARY,  1908. 

WILGUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  &  H.  R.  R.  R. 


Papers.]   ELECTRIFICATION,  SUBURBAN  ZONE,  N.    V.  C.  &  11.  H.  R.  R.        93 


ffi 

fCO 

r-t 

andl 
tion, 
char 
tena 

®  (ti  p. 
■CO  (t 

o 

S-S3 

E 

ing 
inc 
■ges 
nee 
ctur 

»| 

»! 

and 
ludm 
and 
of  la 
es 

o 

1 
1 

B      IJ<?  -• 

P  P  x^ 

1 

&5-£9 

1 

Total 

Mott 
day 

4J4%  on 
5%  on  $ 
General 
Runninj 
Trips  to 
Use  of  s 

1 

1 

"a               ? 

r  .335  da 
aven  e 

at  We 

g  at  Mc 
1  shops. 

-    a                v: 

:  g2S 

if       1 

■   "^Kt 

-1 

H 

:   o              ps 

:  ^Sa 

o' 

.  e              cr 

75   '^"-fl 

Xi 

:  Jg              5" 

;•£        ? 

a: 

.      P5                             to 

H 

.     B                     01 

H 

•    rt-                    '^ 

>■ 

:  -               <_ 

S 

■     CO                       o' 

■      05                            (t 

m 

^ji^^ 

> 

COOi  l-^  *J 

p3 

O00  4i  O 

B  O 

0  d 

^ 

^ 

e  B 

>4^ 

1-1                                  CO 

^              ^ 

K 

^ 

iffl                          to 

OO                     ~>03 

ca                       to 

>(>.                  cnco 

o 

»i                     O-i 

bi 

o                     bi 

s       s'g 

o 

o                       c 

CO                            p 

l5^ 

o 

CO                                  C! 

V-  f^ 

-J                   *^ 

w 

<».;;•                  H 

5s-                 S 

o^g2  2o^ 

a  to                    ^ 

cct;-             o 

aiJJ                    CO     ► 

?e!;  p  o«« 

3  O  ^       -  o 

c^«            s 

^^ffij.^s 

S^5 

=5 

O 

(T> 
CO 

•  <i          p 

3§P^ 

O 

;  P              <! 

^Ci 

•5' 

•     W                     P 

ft  re 

~ 

:  »             p 

P 
3  ' 

3 

.   01              c 

5 

:  B.             X 

C- 

r      1 

:  5-             § 

B 

:1        1 

P  ■ 

W 

(T>  . 

p 

a- 

;  s       8  : 

2.: 

1 

p; 

1 

_^ 

t>* 

^sg| 

B  o 

^ 

^ 

^ 

C  B 

Oi 

CO 

t-1  >— . 

3  <^ 

OS 

to                             ■** 

<I               into 

."•d     ; 

3 

S                2 

£          SS 

% 

P 

P 

*4^ 

tn                          to 

£   1 

CJ> 

|C>. 

td 


c 


!^ 

o 

H 

M 

r, 

W 

*] 

w 

m 

H 

Q 
o 

m 

b 

H 
r/3 

w 

in 

»« 

1:1 

> 

> 

H 

Kl 

i 

§ 

r 

y  GO 

%  5 

O  O 

«^ 

?  o 

H 
3: 

O 

M 
o 
o 
o 


94        ELECTKIFICATIOX,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.   [Papers. 

motive  are  higher  than  those  of  the  steam,  owing  to  the  greater  first 
cost,  the  net  result  is  in  favor  of  the  electric  locomotive,  due  to  lower 
costs  for  repairs  and  maintenance.  These  results  are  based  on  actual 
observations  of  the  steam  locomotive  covering  a  period  of  several 
years;  and  of  the  electric  locomotive  for  two  years  on  the  experi- 
mental track  near  Schenectady  and  one  year  in  the  New  York  zone. 
The  reasons  for  the  lower  cost  of  repairs  on  the  electric  machine  are 
the  simplicity  of  construction  and  the  minimum  number  of  mechani- 
cal parts.  It  is  also  worthy  of  comment  that  the  electric  locomotive 
costs  very  much  less  per  day  for  repairs  and  maintenance,  due  to  lower 
expenses  for  land  and  structures,  and  fewer  days  out  of  service.  For 
instance,  the  fixed  charges  and  cost  of  maintenance  and  operation 
of  the  extensive  steam  engine  plant  on  costly  land,  are  comparable  with 
the  simple  inspection-shed  charges  of  the  electric  locomotive. 

The  Schenectady  experiments  indicated  that  the  cost  of  repairs  of 
the  electric  locomotive  of  this  type  is  about  two-fifths  of  that  of  the 
steam  locomotive  of  a  corresponding  age  and  capacity. 

The  results  of  these  observations  are  shown  in  detail  in  Table  2, 
Plate  XXI,  and  are  summarized  in  Table  3.  They  show  that,  under 
the  stated  conditions,  the  electric  locomotive  has  the  following  ad- 
vantages over  its  steam  rival : 

19%  saving  in  locomotive  repairs  and  fixed  charges. 
18%  saving  in  dead  time  for  repairs  and  inspection. 
25%  greater  daily  ton-mileage. 

6%  saving  in  locomotive  ton-mileage  in  hauling  service. 
11%  saving  in  locomotive  ton-mileage  in  switching  service. 
16%  saving  in  locomotive  ton-mileage  in  road  service. 
12%  net  saving  in  cost    in  hauling  service. 
21%  net  saving  in  cost  in  switching  service. 
2Y%  net  saving  in  cost  in  road  service. 

Even  better  results  may  be  expected  during  winter  months,  when 
steam  locomotives  are  subjected  to  many  conditions  that  cause  addi- 
tional expenses  not  incident  to  the  electric  locomotive. 

Reduced  Cost  of  Grand  Central  Terminal  Operation. — Owing  to 
the  partial  use  of  steam  switching  locomotives,  and  the  presence  of 
the  New  Haven  Company's  steam  road  locomotives  at  the  terminal, 
the  full  benefits  of  change  of  motive  power  have  not  yet  been  secured. 


PLATE  XXI. 

PAPERS  AM.  300.  C.   E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  i  H.  R.  R.  R. 


?tv.S£f.'^isf.v&^."' 


96       ELECTRIFICATION,  SUBURBAN  ZONE,  N.  Y.  C.  &  H.  R.  R.  R.   [Papers. 

However,  on  the  same  wage  basis  for  1907  as  for  1906,  the  month  of 
August,  1907,  showed  a  decrease  in  cost  of  terminal  locomotive  and 
yard  operation  of  nearly  $3  000,  although  the  number  of  cars  in  and 
out  increased  from  64  984  to  68  519.  In  other  words,  the  cost  of  opera- 
tion decreased  9%  while  the  work  done  increased  5^%,  which  is 
equivalent  to  a  net  saving  of  13  J  per  cent. 

Increased  Revenue. — As  to  the  fourth  expectation — increased 
revenue  from  a  larger  volume  of  business — no  definite  conclusions 
can  be  reached  until  the  extension  of  electrical  service  and  the  com- 
pletion of  the  various  other  improvements  afford  an  opportunity  for 
increase  in  frequency  and  speed  of  train  service;  for  the  production 
of  revenue  from  various  sources  at  the  terminal;  and  for  the  expansion 
of  business  that  is  sure  to  follow  the  enlargement  of  the  facilities  of 
the  company  throughout  the  suburban  zone,  not  only  as  regards  the 
local  service,  but  in  an  even  larger  degree  from  long-haul  freight  and 
passenger  traffic. 

Summary  of  Results. — To  summarize,  the  observations  thus  far 
made  demonstrate  that  this  pioneer  electric  installation  in  heavy- 
traction  trunk-line  work  in  the  United  States  has  fully  accomplished 
the  purposes  that  prompted  its  adoption,  namely : 

(1).— Abolition  of  nuisances  incident  to  the  steam  locomotive; 
and 

(2). — Increased  capacity  of  the  Grand  Central  Terminal, 
a  full  year  in  advance  of  the  date  fixed  by  law;  and  in  addition: 

(3). — The  promise,  with  the  completion  of  the  changes,  of  a 
saving,  in  cost  of  operation,  of  from  12  to  27%,  after  pro- 
viding for  increased  capital  charges  for  electrification; 
and 

(4). — The  outlook  of  a  large  future  growth  of  remunerative 
traffic,  and  other  sources  of  revenue  attendant  on  the  use 
of  electricity,  much  more  than  sufficient  to  provide  for  the 
increased  capital  charges  for  the  other  improvements. 

Several  years  w;ill  be  consumed  in  the  gradual  rounding  out  of  the 
work  as  a  whole;  but  it  is  gratifying  to  have  this  early  indication  of 
the  success  of  the  undertaking  from  both  the  engineering  and  financial 
standpoints. 


PLATE  XXII. 

PAPERS,  AM.  80C.  C.  E. 

FEBRUARY,  1908. 

WILQUS  ON 

ELECTRIFICATION  OF  SUBURBAN  ZONE 

OF  N.  Y.  C.  &  H.  R.  R.  R. 


AUgmnent. 

Highway  (No's).     J^^ 
Geography. 


♦ 

Foldout 

Here 
♦  ♦ 

♦ 


I 


Sub.  Loop 


"Grades 


Vol.  XXXIV.  FEBRUARY,   1908.  No.   2. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEEES. 

INSTITUTED     18. t  2. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


THE  FLOOD  OP  MARCH,  1907, 

IN  THE  SACRAMENTO  AND  SAN  JOAQUIN  RIVER 

BASINS,  CALIFORNIA.*! 


By  W.  B.  Clapp,  M.  Am.  Soc.  C.  E.,  E.  C.  Murphy,  Assoc.  M.  Am. 
Soc.  C.  E.,  AND  W.  F.  Martin,  Jun.  Am.  Soc.  C.  E. 


Introduction. 


The  Sacramento  and  San  Joaquin  Valleys  were  visited,  in  March, 
1907,  by  one  of  the  most  destructive  floods  that  have  ever  occurred  in 
California,  the  resulting  financial  loss  being  unquestionably  greater 
than  that  from  any  other  flood  of  which  there  is  record.  The  greatest 
damage  was  done  in  the  valleys  of  the  trunk  streams,  especially  Sacra- 
mento Valley.  The  Lower  Sacramento  Eiver  and  its  two  largest 
tributaries.  Feather  and  American  Elvers,  reached  the  highest  stages 
ever  recorded,  and  record  stages  were  reached  by  other  tributaries  of 
the  Sacramento  and  by  the  San  Joaquin  and  its  tributaries. 

The  flood  was  remarkable  in  many  respects.  In  the  first  place,  it 
was  preceded  by  a  period  of  heavy  precipitation,  and  consequent  flood 
stages  of  all  the  streams,  a  condition  which  had  prevailed  intermittently 
for  several  preceding  weeks.     As  a  result,  the  earth  was  thoroughly 

*The  data  upon  which  this  paper  is  based  were  collected  by  the  VVater  Resources 
Branch  of  the  United  States  Geological  Survey  in  co-operation  with  the  State  ot  Calitornia, 
and  the  naner  is  nublishel  bv  oermission  of  the  Director  ot  the  Survey. 

Further  acknowledgments  are  due  to  Mr.  J.  H.  Scarr,  the  district  forecaster 
of  the  United  States  Weather  Bureau,  and  to  the  engineering  department  of  the 
Southern   Pacific  Railway,    for  data   furnished. 

tThis  paper  will  not  be  presented  at  any  meeting,  but  written  communications  on  the 
subject  are  invited  foi  publication  with  it  in  Transactions. 


100  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

saturated,  and  all  the  surface  basins  which  impound  and  store  flood 
waters  temporarily  were  full.  Particularly  was  this  true  of  the  large 
flood  basins  on  each  side  of  the  Sacramento  Eiver.  Then,  too,  this 
flood  was  due  to  a  general  precipitation  of  extraordinary  intensity 
throughout  the  entire  drainage  basin  (the  storm  covering  a  period  of 
several  consecutive  days),  and  also  to  comparatively  high  tempera- 
ture and  consequent  rapid  melting  of  snow  in  the  higher  altitudes. 

This  flood  was  remarkable,  also,  because  of  the  record-breaking 
stages  of  so  many  of  the  streams,  such  as  the  Lower  Sacramento  River 
and  the  Feather,  Yuba,  and  American  Rivers.  Not  only  were  they 
higher  than  ever  known  before,  but  they  maintained  their  high  stages 
for  a  moderately  long  period.  All  the  other  streams  of  the  water-shed 
also  maintained  high  stages  for  a  like  period,  so  that  the  resultant  was 
a  flood  of  exceptional  height  and  extent,  and  of  considerable  duration. 
For  the  4-day  period,  March  18th  to  21st,  the  mean  rate  of  run-oif 
from  the  mountains  and  foot-hills  of  the  Sacramento  Basin  alone  was 
about  530  000  cu.  ft.  per  sec,  or  more  than  22  cu.  ft.  per  sec.  per  sq. 
mile. 

During  this  flood,  special  effort  was  made  by  the  engineers  of  the 
United  States  Geological  Survey  to  obtain  valuable  flood  data.  The 
flow  of  nearly  all  the  important  tributaries  of  both  the  Sacramento  and 
San  Joaquin  River  systems  was  gauged  in  the  foot-hills  above  the  point 
of  debouchure.  The  flow  from  83%  of  the  mountains  and  foot-hills 
in  the  Sacramento  Basin  was  measured  at  eleven  gauging  stations. 
In  the  San  Joaquin  Basin  the  flow  from  41%  of  the  mountains  and 
foot-hills  was  measured  at  six  gauging  stations.  Unfortunately,  no 
gaugings  were  made  of  the  San  Joaquin  itself. 

It  is  believed  that  the  data  obtained  during  this  flood  will  fully 
repay  the  State  of  California  for  its  generous  co-operation  with  the 
United  States  Geological  Survey  in  the  study  of  its  water  resources. 
Data  are  now  available  for  planning  for  these  great  valleys  a  more 
comprehensive  reclamation  system  than  has  been  possible  heretofore. 
The  importance  of  the  data  collected  will  be  appreciated  when  it  is 
recalled  that  the  rate  of  run-off  from  the  mountains  and  foot-hills  of 
the  Sacramento  Basin  alone  for  a  period  of  4  consecutive  days,  March 
18th  to  21st,  was  112%  greater  than  the  rate  used  as  a  maximum  by 
the  1904  Commission  of  Engineers,  after  a  careful  study  of  all  flood 
data  on  record,  including  those  of  the  1904  flood.     It  is  doubtful  if 


Papers.]    THE  FLOOD  0¥  MARCH,  1907,  IN  CALIFORNIA  RIVERS  101 

any  combination  of  causes  or  conditions  will  ever  produce  a  larger 
rate  of  delivery  of  water  to  this  valley  for  a  4-day  period  than  occurred 
during  the  flood  of  March,  1907. 

Topography  and  Drainage  of  the  Water-shed. 

California  is  traversed,  in  a  general  northwest-southeast  direction, 
by  two  distinct  and  approximately  parallel  ranges  of  mountains  which 
extend  almost  the  entire  length  of  the  State.  Near  the  eastern  border 
is  the  Sierra  Nevada;  not  far  from  the  shore  line  on  the  west  is  the 
Coast  Kange.  These  two  ranges  merge  into  each  other  about  40  miles 
south  of  the  California-Oregon  boundary  line,  the  meeting  point  being 
Mount  Shasta,  which  has  an  elevation  of  14  380  ft.  They  are  merged 
again  south  of  Bakersfield  by  a  cross-range  known  as  Tehachapi 
Mountains. 

The  elevation  of  the  Sierra  Nevada  ranges  from  about  6  000  ft. 
east  of  Mount  Shasta  at  the  north,  to  14  501  ft.  south  of  Yosemite 
National  Park  where  the  range  culminates  in  Mount  Whitney. 
Beckwith  Pass,  about  150  miles  south  of  the  northern  boundary 
line,  is  the  lowest  pass  through  the  range,  and  has  an  elevation  of 
5  300  ft.  The  Coast  Range  is  comparatively  low,  and  is  unbroken  ex- 
cept at  Carquinez  Strait  and  the  Golden  Gate  which  permit  the  drain- 
age through  Suisun  Bay  to  reach  the  Pacific. 

The  Sierra  Nevada  and  Coast  Ranges,  merging  at  the  north  and 
south,  inclose  a  water-shed  approximately  58  000  sq.  miles  in  area, 
with  a  single  outlet  near  the  middle  of  the  western  side.  This  water- 
shed is  somewhat  elliptical  in  shape,  and  has  a  length  of  about  540 
miles  from  north  to  south  and  a  width  varying  from  120  to  150  miles. 
It  is  drained  by  two  large  river  systems,  the  Sacramento  in  the  north 
and  the  San  Joaquin  in  the  south,  and  these  are  quite  commonly  re- 
ferred to  as  the  Sacramento  and  the  San  Joaquin  River  Basins. 

Sacramento  River  has  its  source  in  the  region  of  Mount  Shasta, 
and  flows  almost  due  south  through  the  trough  of  the  water-shed  until 
it  discharges  into  Suisun  Bay.  The  San  Joaquin  rises  in  the  Sierra 
Nevada,  in  the  region  of  Mount  Lyell,  just  east  of  Yosemite  National 
Park,  at  an  elevation  of  13  000  ft.,  and  flows  southwestward  until  it 
emerges  from  the  foot-hills  into  the  trough  of  the  valley,  when  it  turns 
and  flows  northwestward  to  its  junction  with  the  Sacramento  near 
Suisun  Bay,  through  which  the  combined  volume  of  the  two  systems 


102  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

finds  an  outlet  to  the  Pacific  by  way  of  San  Pablo  and  San  Francisco 
Bays  and  the  Golden  Gate. 

The  drainage  of  the  water-shed  determines  its  division  into  three 
distinct  basins :  On  the  north  is  the  Sacramento  Basin,  27 100  sq. 
miles  in  area,  drained  by  the  Sacramento  River  and  its  tributaries;  in 
the  center  is  the  San  Joaquin  Basin,  about  18  300  sq.  miles  in  area, 
drained  by  the  San  Joaquin  and  its  tributaries  (excluding  Kings 
Eiver,  which,  for  reasons  given  later,  is  classified  under  the  Lake 
Basin) ;  in  the  south  is  the  Lake  Basin,  with  an  area  of  about  12  600 
sq.  miles,  containing  several  lakes  with  their  tributary  drainage,  but 
at  the  present  time  having  no  outlet  discharging  to  the  sea. 

That  portion  of  the  three  basins  which  is  inclosed  by  the  sharply- 
defined  line  of  the  foot-hills  is  called  the  "Great  Valley  of  California." 
This  valley  has  a  length  of  about  400  miles  from  north  to  south,  an 
average  width  of  about  40  miles,  and  an  area  of  probably  15  000  sq. 
miles,  and  is  surrounded  by  steep  mountains.  The  western  mountain 
slope — that  of  the  Coast  Range — is  comparatively  narrow,  having  an 
average  width  of  about  18  miles.  Considering  the  entire  length  of 
the  district,  from  north  to  south,  the  precipitation,  as  a  whole,  is  light, 
and  perennial  streams  are  few,  but,  in  the  region  about  Clear  Lake 
and  Mount  St.  Helena,  in  the  Lower  Sacramento  Basin,  the  precipita- 
tion is  remarkably  heavy,  and  occurs  almost  entirely  as  rain.  The 
eastern  mountain  slope,  which  has  an  average  width  of  about  58  miles, 
is  visited  by  rather  heavy  precipitation  throughout  almost  its  entire 
length  from  north  to  south,  particularly  in  the  central  part  of  the 
Sacramento  Basin.  A  large  percentage  of  the  precipitation  occurs  as 
snow  on  the  higher  elevations.  From  this  slope  come  all  the  larger 
tributaries  to  the  Sacramento  and  San  Joaquin  Rivers  and  the  San 
Joaquin  itself,  as  well  as  the  principal  tributaries  to  the  Lake  Basin. 
The  change  from  mountain  to  valley  is  quite  abrupt  along  a  well- 
defined  line,  but  the  slope  of  the  valley  is  gentle  and  uniform. 

What  is  commonly  called  the  Sacramento  Valley  extends  north- 
ward only  to  Iron  Canyon,  near  Red  Bluff.  In  the  Report  of  the  Conl- 
missioner  of  Public  Works  to  the  Governor  of  California,  in  1894 
(page  28),  the  valley  is  described  as  having  a  total  area  of  about  4  250 
sq.  miles,  divided  as  follows :  2  510  sq.  miles  of  high  lands,  not  sub- 
ject to  overflow;  450  sq.  miles  of  lower  lands,  overflowed  occasionally 
by  high  floods;   1250  sq.  miles  of  low  lands,  overflowed  periodically; 


Papers.]    THE  FLOOD  OF  .MARCH,  1907,  IN  CALIFORNIA  RIVERS 


103 


d:. 


4      .^-^'  ¥■ 


ruylcL.      Jw 


MS 


.^  ICei. 
16   f- 


LEGEND 

+  tT.s.  Gauging  Stations 
•    U.S.  Precipitation  Stations 
"V  l'roi)Osecl  Reservoir  Sites 
f-'  Boundary  of  Watcr-slied 
Scale  of  Miles 

20       0       20       iO       «0       80 
190T 


San  Franciscb, 


MAP  OF  WATER-SHED 

OF 

SACRAMENTO  AND  SAN  JOAQUIN   RIVERS 

IN  CALIFORNIA 

SHOWING 

TRIBUTARV  STREAMS,     UNITED   STATES  GAUGING  AND 

PRECIPITATION  STATIONS, 

AND  PROPOSED  RESERVOIR  SITES 

IN  SACRAMENTO  BASIN 


104  THE  PLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

and  38  sq.  miles  of  perennial  stream  surface.  Below  the  mouth  of 
Stony  Creek  (Plate  XXIli  and  Fig.  1)  the  central  portion  of  the 
valley  is  a  flood  piano  of  unusual  extent,  the  immediate  river  banks 
being  from  5  to  20  ft.  higher  than  the  land  on  either  side  some  distance 
from  the  river.  In  the  vicinity  of  the  river  banks  the  ground  slopes 
rapidly  from  the  river  toward  the  trough  of  the  flood  basins  on  either 
side,  but,  as  the  bottom  of  the  trough  is  approached,  the  slope  becomes 
more  gradual.  The  lowest  portions  of  the  flood-basin  troughs  are  from 
2  to  7  miles  from  the  river  channel. 

The  large  flood  basin  on  the  west  side  of  the  Sacramento  is  divided 
into  two  smaller  basins  by  a  ridge  of  debris  brought  down  by  Cache 
Creek.  These  are  the  Colusa  Basin  in  the  north  and  the  Yolo  Basin  in 
the  south.  The  large  flood  basin  on  the  east  side  of  the  Sacramento 
is  divided  into  foiir  smaller  basins  by  Marysville  Buttes  and  the  Feather 
and  American  Eivers.  From  north  to  south,  they  are  called  Butte 
Basin,  Sutter  Basin,  American  Flood  Basin,  and  Sacramento  Flood 
Basin.  Fig.  1  shows  the  position  of  the  flood  basins.  The  following 
data  regarding  the  area  and  capacity  of  these  smaller  flood  basins  are 
taken  from  the  Report  of  the  Commissioner  of  Publi<3  Works  to  the 
Governor  of  California,  for  1894: 

Colusa  Basin  is  50  miles  long,  from  2  to  7  miles  wide,  and  has  a 
capacity  of  690  000  acre-ft.  at  flood  stage.  It  discharges  into  Sacra- 
mento River  above  Knight's  Landing  through  Sycamore  Slough. 

Yolo  Basin  has  a  length  of  40  miles,  an  average  width  of  7  miles, 
and  a  capacity  of  1 115  000  acre-ft.  at  flood  stage.  It  discharges 
through  Cache  Slough  into  Steamboat  Slough,  and  thence  into  the 
Sacramento  near  the  foot  of  Grand  Island,  about  25  miles  above  the 
head  of  Suisun  Bay. 

Butte  Basin  is  north  of  Marysville  Buttes,  and  has  an  area  of 
from  30  to  150  sq.  miles,  depending  upon  the  river  stage,  and  a  capacity 
cf  460  000  acre-ft.  at  flood  stage.  It  discharges  through  Butte  Slough 
into  Sutter  Basin. 

Sutter  Basin  is  south  of  Marysville  Buttes  and  north  of  the  Feather 
River.  It  has  an  area  of  138  sq.  miles,  and  a  capacity  of  895  000  acre- 
ft.  at  flood  stage.  It  discharges  through  sloughs  into  Sacramento 
River  above  the  mouth  of  Feather  River. 

The  American  Flood  Basin  is  south  of  Feather  River  and  north  of 
the  American.     It  has  an   area   of   110  sq.   miles,   and  a  capacity  of 


PLATE  XXIII. 

PAPERS,   AM,  30C.   C.    E. 

FEBRUARY,  1908. 

OLAPP,  MURPHY  AND  MARTIN  ON 

FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS. 


MAP  OF 
THE  LOWER  PORTION  OF  THE' 

SACRAMENTO  AND  SAN  JOAQUIN 
VALLEYS 

STATE  OF  CALIFORNIA 

SHOWINQ 

APPROXIMATE  CONTOURS   REFERRED  TO  MEAN   SEA  LEVEL. 

FLOOD  BASINS  IN   SACRAMENTO  VALLEY,  AND 

THE  MARGIN  OF  THE   FLOOD  PLAIN  OF 

MARCH.    1907 

6  0  5  I'o  Ifi 

NOTK:  The  mai^ln  of  the  flouil  plane  abown  thus,  [iwj;iiiU,tj 
does  not  neceasarily  indicate  the  area  actually  covered  by  water, 
but  It  shows  approximately  the  area  which  was  below  the  \ 
surface  In  the  river.. 


] 


♦ 

Foldout 
w  Here 


p" 


♦ 


♦ 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  107 

Climate. 

The  climate  of  California  is  probably  one  of  its  most  valuable 
assets.  The  principal  factors  affecting  the  climate  are  proximity  to 
the  Pacific  Ocean,  and  diversified  topography.  The  warm  Japanese 
ocean  currents,  which  bathe  about  1  000  miles  of  the  coast  line,  serve 
to  equalize  the  temperature  as  normally  affected  both  by  seasons  and 
latitude.  The  influence  of  the  topography  is  such  that  altitude  rather 
than  latitude  is  the  chief  factor  affecting  temperature. 

As  regards  precipitation,  the  year  is  divided  into  two  well-defined 
seasons,  the  "rainy  season"  from  November  to  March,  and  the  "dry 
season"  from  April  to  October.  The  rainy  season  is  usually  marked 
by  a  series  of  storms,  of  greater  or  less  severity,  which  form  in  the 
Pacific  Ocean  and  move  eastward  to  the  coast,  depositing  their  moisture 
before  crossing  the  Sierra  Nevada.  The  centers  of  the  most  severe 
storms  generally  strike  the  coast  in  the  State  of  Washington  and  then 
move  southward  through  Oregon  into  California  between  the  mountain 
ranges.  These  storms  almost  invariably  make  their  appearance  in  late 
winter  or  early  spring,  being,  as  a  rule,  most  severe  about  the  time  of 
the  vernal  equinox.  At  this  season  the  precipitation  is  quite  general 
throughout  the  State,  increasing  with  altitude  and  also  with  latitude. 

Flood  Conditions  and  Causes. 

During  the  winter  and  early  spring  of  each  year,  toward  the  end 
of  the  rainy  season,  the  various  streams  of  the  Sacramento  and  San 
Joaquin  Basins  generally  reach  their  highest  stages.  The  most  serious 
flood  conditions  invariably  exist  on  the  lower  courses  of  the  trunk 
streams,  the  Sacramento  and  the  San  Joaquin.  On  the  Sacramento 
River,  in  particular,  serious  damage  is  inflicted  on  crops  and  trans- 
portation interests  almost  every  year.  Of  course,  the  destructiveness 
of  any  flood  is  measured  largely  by  its  height  and  duration.  In  these 
basins  the  maximum  height,  and  generally  the  greatest  duration,  of 
floods  on  the  primary  streams  result  from  the  simultaneous  flooding 
of  all  the  secondary  and  tertiary  streams,  a  condition  which  obtains 
when  there  is  a  period  of  long-sustained  precipitation  throughout  the 
entire  water-shed,  accompanied  by  high  temperature  and  rapid  melting 
of  snow  on  the  higher  elevations.  It  was  such  a  condition  that  brought 
about  the  flood  of  March,  1907. 


108  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

Other  conditions  that  contribute  more  or  less  to  all  floods  in  this 
area  are  the  following : 

1. — The  steep,  barren,  and  impervious  slopes  of  the  mountains  and 
foot-hills,  which  result  in  streams  of  heavy  grades  and  the  rapid  de- 
livery of  water  to  the  valleys. 

2. — The  broad,  flat  valleys,  with  light  grades  and  sluggish  streams. 

3. — The  limited  channel  capacity.  It  is  said  that  some  of  the 
trunk  channels  are  not  large  enough  to  carry  even  one-third  of  the 
flood  flow.  Particularly  is  this  true  of  the  Sacramento  River.  Here 
the  surplus  water  overflows  into  the  flood  basins,  the  result  being 
either  to  increase  or  diminish  the  stage  of  the  lower  course  of  the  river, 
depending  on  the  volume  of  water  in  the  flood  basins  at  the  beginning 
of  the  flood  period  and  the  duration  of  the  period. 

4. — The  common  outlet  of  the  two  river  systems,  with  large  tribu« 
taries  of  each  system  discharging  into  trunk  streams  near  this  outlet. 

5. — The  constriction  of  the  flood  area  in  the  delta  of  the  two  rivers 
through  the  reclamation  of  large  areas  of  overflow  land  by  levees. 

6. — The  deposition  of  the  debris  resulting  from  hydraulic  mining 
in  several  tributaries  of  the  Sacramento  River,  the  result  of  which  has 
been  the  filling  of  channels  and  the  reduction  of  gradients,  there- 
by raising  the  flood  plane  several  feet. 

7. — The  tidal  and  wind  action  in  the  delta  of  the  two  rivers. 

Precipitation. 

In  the  Sacramento  Valley,  the  mean  annual  precipitation  varies 
from  15  in.  in  the  southern  to  20  in.  in  the  northern  part,  while,  in  the 
tributary  foot-hill  and  mountain  areas,  it  varies  from  20  to  60  in., 
with  an  occasional  maximum  of  100  in.  In  the  San  Joaquin  Valley, 
the  mean  annual  precipitation  varies  from  10  in.  in  the  southern  to  15 
in.  in  the  northern  part,  and  in  the  foot-hill  and  mountain  areas  it 
varies  from  15  to  40  in.  In  the  Sierras,  the  greater  part  of  the  pre- 
cipitation is  normally  in  the  form  of  snow,  and  the  magnitude  of 
floods  depends  largely  on  its  rate  of  melting.  A  heavy,  warm  rain  on 
a  deep,  freshly  fallen  snow  produces  a  maximum  run-off. 

In  January  and  February,  1907,  there  were  two  periods  of  heavy 
and  long-sustained  precipitation,  one  from  January  2d  to  17th,  and 
the  other  from  January  24th  to  February  4tli.  The  precipitation  was 
unusually  heavy  over  the   Sacramento  Basin,  diminishing  gradually 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  109 

toward  the  south.  The  precipitation  during  the  first  of  these  periods 
produced  the  ordinary  winter  stages  of  the  tributaries  of  the  Sacra- 
rcento  River;  that  during  the  second  period  produced  flood  stages  on 
the  tributaries  of  both  the  Sacramento  and  San  Joaquin  Rivers  and 
high  stages  in  the  Lower  San  Joaquin.  American  and  Bear  Rivers 
reached  stages  almost  as  high  as  in  the  great  flood  of  the  following 
month.  Yuba  River  was  higher  than  at  any  time  previously  re- 
corded. 

In  March  there  were  two  precipitation  periods,  one  from  the  2d  to 
the  11th,  in  which  the  amount  of  rainfall  was  moderate,  and  the  other 
from  the  16th  to  the  25th,  in  which  it  was  extraordinarily  heavy.  The 
precipitation  of  the  latest  period  was  accompanied  by  unusually  high 
mean  temperatures,  especially  in  the  higher  altitudes,  from  the 
Feather  River  south  to  the  Tuolumne,  causing  very  rapid  melting  of 
snow  and  exceedingly  large  run-ofl.  The  average  from  24  fairly  rep- 
resentative meteorologic  stations  throughout  the  basin  shows  that  the 
mean  temperature  for  March  17th  to  20th  was  about  5°  above  the  mean 
for  the  month,  with  low  daily  maxima  resulting  from  cloudiness  and 
rain,  and  high  daily  minima  due  to  the  liberation  of  heat  by  the 
storm.  The  average  greatest  daily  range  in  this  period  was  only  16 
degrees.  These  facts  indicate  that  probably  all  stations  with  a  monthly 
mean  temperature  as  high  as  25°  had  scarcely  any  freezing  condi- 
tions from  March  16th  to  20th,  when  the  precipitation  was  heaviest. 
Further,  they  show  that,  out  of  113  stations  located  at  various  eleva- 
tions throughout  the  Sacramento  and  San  Joaquin  drainage  basins, 
at  105  of  them  all  the  precipitation  from  March  17th  to  20th  was  prob- 
ably in  the  form  of  rain  or  of  snow  in  a  melting  condition. 

Table  1  shows  the  monthly  precipitation  from  January  to  March, 
1907,  the  daily  precipitation  for  the  three  days,  March  17th,  18th,  and 
19th,  when  it  was  greatest,  and  the  precipitation  for  the  ten  days, 
March  17th  to  26th,  for  120  places  in  the  Sacramento  and  San  Joaquin 
Basins,  varying  in  altitude  from  20  to  7  017  ft.,  arranged  according  to 
basins  of  tributary  streams.  Where  possible,  the  monthly  precipita- 
tion for  the  period  of  January  to  March,  1904,  is  also  given,  for  com- 
parison with  the  great  flood  of  that  year. 

Table  2  shows  the  average  precipitation  and  average  mean  tempera- 
tures for  March  at  113  'stations  arranged  according  to  stream  basins 
and  in  order  of  altitude. 


110  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

Table  3  shows  the  results  at  24  stations  ranging  in  altitude  from 
60  to  5  270  ft.,  the  data  having  been  taken  from  the  Climatological 
Report  of  the  United  States  Weather  Bureau  for  March,  1907.  These 
particular  stations  were  selected  because  they  are  the  only  ones  in  the 
basin  for  which  daily  temperatures  have  been  published.  They  are 
fairly  well  distributed,  both  as  regards  area  and  altitude,  and  are  prob- 
ably as  representative  as  any  that  could  have  been  chosen.  This  table 
also  shows  the  extraordinary  intensity  of  precipitation  from  March 
17th  to  26th  by  percentages  with  reference  to  the  total  for  the  month, 
and  also  the  normal  for  the  month,  covering  a  period  of  21  years  on 
an  average. 

These  tables  show  conclusively  that  the  precipitation  from  March 
17th  to  26th,  and  particularly  on  March  17th,  18th,  and  19th,  was 
phenomenally  heavy  for  this  section  of  the  country.  This  large  pre- 
cipitation is  rather  evenly  distributed  throughout  all  the  river  basins, 
but  there  is  a  very  noticeable  and  quite  rapid,  though  comparatively 
regular,  increase  with  the  altitude.  During  the  month,  sixteen  stations, 
with  elevations  of  more  than  3  500  ft.,  had  more  than  30  in.  in  depth 
of  precipitation;   about  forty  stations,  with  elevations  of  more  than 

1  500  ft.,  had  more  than  20  in. ;  and  fully  one-third  of  the  total  pre- 
cipitation for  the  month  fell  on  March  17th,  18th,  and  19th.  On  one 
of  these  three  days,  seventeen  stations,  with  an  altitude  of  more  than 

2  000  ft.,  had  precipitations  of  from  5  to  8  in.  in  24  hours.  It  is  note- 
worthy that  the  range  of  temperature  with  altitude  was  quite  regular, 
and  that  there  were  no  very  low  temperatvires  even  at  very  high  eleva- 
tions. It  is  highly  probable  that  at  elevations  of  5  000  ft.  a  large  part 
of  the  precipitation  occurred  as  rain  or  as  snow  which  melted  rapidly. 
Indeed,  at  Inskip,  in  the  Feather  Basin,  with  an  elevation  of  4  850 
ft.,  a  24-hour  rainfall  of  8  in.  was  reported.  Taking  a  record  of  21 
years  on  an  average  throughout  the  basin,  it  is  seen  that  about  88%  of 
the  normal  precipitation  for  March  occurred  on  March  17th,  ISth,  and 
19th,  1907,  or,  counting  20  days  as  normally  rainy  in  this  month,  the 
intensity  of  this  3-day  period  was  about  600%  of  the  normal  intensity 
for  the  month.  During  these  3  days  the  average  precipitation  at  the 
sixteen  stations,  principally  in  the  Feather  and  Yuba  Basins,  having 
more  than  30  in.  during  the  month,  was  145%  of  the  normal  for  the 
month,  or  at  an  average  intensity  of  1  000%  of  the  normal. 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  111 

TABLE  1. — Precipitation. 


Rainfall  IStatiosts. 

Precipitation, 

[N  Inches. 

g^j 

2.^ 

a 

m 

A 

M 

^ 

.n 

a^ 

No. 

Name. 

a 

oi 

■-5 

^ 

3 

3fl 

t~ 

QO 

OS 

m 

"  o 

S 

fi 

fa 

g 

o  o 

s 

^ 

^5 

S£ 

W 

M 

S 

a 

g 

^-d 

Sacramento  Drainage  Rasin. 


1 

Delta 

1138 

1904 
1907 

3.96 
12.26 

21.19 
9.21 

23.93 
24.45 

49.08 
45.92 

1.10 

3.50 

3.50 

18.10 

47.3= 

2 

Redding 

565 

1S04 
1907 

2.34 

8. '57 

14.10 
9.09 

15.89 
7.28 

32.83 
24.94 

1.36 

0.88 

0.11 

4.57 

49.0^ 

a 

Dunsmuir 

2  285 

1904 
1907 

5.03 
20.53 

24.00 
8.27 

22.90 
18.64 

51.93 
47.44 

2.00 

2.87 

2.72 

13.61 

46.9° 

4 

SissoD 

3  555 

1904 
1907 

3.26 
9.48 

10.91 
2.84 

15.90 
13.16 

30.07 
35.48 

0.00 

8.55 

1.27 

11.27 

37.6° 

5 

Nimshew 

2  0(10 

1904 
1907 

10.82 
17.64 

8.72 
13.12 

13.27 
27.69 

32.81 
58.45 

5.54 

4  00 

3.03 

18.36 

42.4° 

fi 

Sacramento 

71 

1904 
1907 

0.15 
4.63 

5.26 
2.37 

5.43 

7.28 

11.14 
14.28 

0.42 

1.74 

0.56 

4.75 

50.9° 

7 

Fruto  

624 

1904 
1907 

0.75 
6.43 

6.13 
1.95 

7.07 
4.67 

13.95 
13.05 

0.45 

0.40 

0.30 

2.90 

50.1° 

8 

Shasta 

1148 

1904 
1907 

2.79 
13.65 

24.86 
7.89 

16.37 
14.47 

44.02 
36.10 

1.54 

3.25 

1.03 

10.98 

48.0° 

9 

Corning 

377 

1904 
1907 

0.60 
3.60 

4.95 
2.60 

7.30 
5.05 

12.85 
11.25 

0.00 

0.95 

0  28 

2.68 

49.1° 

10 

Red  Bluff 

307 

1904 
1907 

1.44 
6.10 

6.63 
3.13 

8.33 
5.92 

16.40 
15.15 

0.62 

0.38 

0.00 

2.93 

48.4° 

11 

Tehama 

220 

1904 
1907 

1.01 
4.75 

4.6r 

2.96 

7.19 
5.38 

12.87 
13.09 

0.86 

0.28 

0.30 

2.':9 

50.8° 

13 

Chico 

189 

1904 
1907 

0.80 
6.28 

5.64 

2.09 

9.33 

8.08 

15.77 
18.67 

1.58 

0.73 

0.36 

4.79 

49.2° 

13 

Durham 

160 

1904 
1907 

1.70 
6.45 

5.75 
2.09 

10.32 
8.39 

17.77 
16.93 

1.61 

0.87 

0.45 

4.49 

50.4° 

14 

Willows 

136 

1904 
1907 

0.45 
4.84 

3.44 
1.02 

7.61 
3.63 

11.50 
9.48 

0.70 

0.13 

0.05 

1.98 

49.1° 

15 

C(jlusa 

60 

1904 
1907 

0.66 
5.63 

3.13 

0.75 

5.67 
3.80 

9.46 
10.18 

0.58 

0.22 

0.00 

2.21 

49.9° 

16 

Suisun 

20 

1904 
1907 

1.12 
8.89 

6.50 
3.59 

7.52 

7.57 

15.14 
20.05 

1.61 

0.08 

1.55 

5.95 

52.0° 

17 

Dunnigan 

65 

1904 
1907 

0.66 
7.63 

5.33 
1.63 

8.87 
6.98 

14.86 
16.24 

1.10 

0.04 

1.40 

3.91 

52.9° 

1H 

Wesi  Branch 

3150 

1907 

17.96 

39.69 

112  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

TABLE  1.— (Continued.) 


Rainfall  Stations. 


Precipitation,  in 

Inches. 

1 

a 

ai 

^' 

xJ 

Jh' 

.a 

-M 

«' 

S 

« 

-g 

^^ 

> 

0) 

Q 

1-5 

ta 

g 

hI 

3 

c6 

S^ 

H 

CO 

S 

^ 

g 

ss 


No. 


Name. 


McCloud  Kiver  Drainage  Basin. 


19   Johns  Camp. 


1904      4.50    19.73    37.36    51.49 


Pit  Kiver  Drainage  Basin. 


^0 

Csdarville 

4  675 
4  460 

1904 
1907 

1907 

1.12 
1.99 

1.35 

4.87 
3.70 

2.87 

4.61 
3.31 

4.13 

10.60 
9.00 

8.35 

0.67 
0.55 

0.40 
0.75 

0.06 
0.14 

1.59 
2.45 

?I1 

Alttiras 

34.4"^ 
35.6° 

Feather  River  Drainage  Basin. 


9f?, 

Magalia 

2  321 

1904 
1907 

3.43 

23.57 

23.39 

10.71 

30.13 
37.75 

56.95 
73.03 

7.65 

6.66 

2.79 

24.48 

41.9° 

9^ 

Oroville.. 

250 

1904 
1907 

1.60 
6.71 

7.99 
3.59 

10.86 
10.90 

20.45 
21.20 

1.10 

1.44 

0.52 

5.57 

51.00 

9A 

Butte  Vallev  

4  030 
3  600 

1904 
1907 

1904 
1907 

4.30 
11.96 

2.39 
9.57 

22.90 
6.78 

18.81 
4.48 

22.10 
26.76 

15.53 
24.51 

49.20 
45.50 

36.73 

38.56 

Greenville 

25 

4.25  6.17 

2.91 

19.89 

37.8* 

9.P, 

4  730 
3  400 

1904 

1904 
1907 

4.13 

2.46 
11.89 

29.10 

23.10 
4.96 

29.90 

10.83 
30.15 

63.13 

35.39 

47.00 

4.40 

25.55 

97 

5.30 

6.50 

35.8° 

•■'8 

4  850 

1907 

45.30 

"9 

Bless                        

98 

1904 
1907 

1.09 
4.55 

4.98 
1.85 

8.35 
6.57 

14.42 
12.97 

0.20 

0.00 

0.00 

3.65 

52.0» 

30 

Brush  Creek 

2  140 

1904 
1907 

4.81 
16.21 

23.11 
11.49 

25.01 
33.02 

58.93 
60.72 

5.70 

5.40 

3.40 

23.96 

42.8° 

31 

67 

1904 
1907 

1.19 
4.52 

5.18 
4.30 

7.77 
10.59 

14.14 
19.41 

0.30 

1.30 

2.00 

6.44 

53.8<> 

391 

213 

1904 
1907 

1.48 
5.86 

7.23 
3.34 

9.35 
8.80 

18.05 
17.50 

0.97 

0.86 

0.25 

4.37 

50.8° 

'« 

Sterling  City      

3  535 

1904 
1907 

3.96 
24.63 

26.51 
17.54 

25.22 
4:^.38 

55.69 
85.55 

6.66 

7.90 

6.16 

32.86 

37.0° 

Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS 
TABLE  1.— {Continued.) 


113 


Rainfall  Stations. 

Precipitation, 

IN  Inches. 

0)  u 

a 

TD 

^' 

.d 

A 

.c 

as 

No 

Name. 

c8 

6 

s 

® 

iL, 

0  o 

CO 

00 

OS 

^^ 

•*^  o 

a-" 

> 

ft 

i-s 

El. 

^' 

e8 

1 

h 
cj 

s 

^i 

SB 

Yuba  Eiver  Drainage  Basin. 


34 

Colgate ^. 

650 
1650 

2  580 

3  200 

5  939 

7  017 

5  000 

3  400 
3  250 

1904 
1907 

19(14 
1907 

1904 
1907 

1904 
1907 

1904 
1907 

1904 
1907 

1904 
1907 

1907 

1907 

1907 

1907 

1907 

1904 
1907 

3.79 

7.86 

3.79 
10.54 

2.76 
10.21 

3.85 
10.25 

5.20 
14.70 

4.20 
13. SO 

4.48 
17.75 

18.75 

12.11 

5.37 
13.82 

9.92 
10.28 

14.04 
8.98 

19.17 

8.22 

16.44 
9.23 

30.80 
6.25 

30.40 

4.38 

30.35 
16.40 

16.78 

12.87 

11.26 

45.61 
12.68 

8.19 
19.31 

13.65 
19. 4S 

18.64 
24.62 

21.89 
28.64 

26.87 
24.20 

21.30 
37.36 

31.66 
42.62 

36.12 

37.38 

36.93 

26.78 

29.01 

39.51 
31.46 

21.90 
37.45 

31.48 
38.95 

40.57 
43.05 

42.18 
48.12 

62.87 
45.15 

55. 9(. 
45.24 

66.49 
76.27 

Dobbins 

35 

2.50 
2.47 
3.07 
1.00 
1.42 
6..58 

2.33  2.60 

3.34  3.63 
4.51  3.97 
1.C0  3.60 
2.42  2.32 
6.19  5.42 

13.04 
17.76 
21.10 
14.10 
16.06 
.33.12 

36 

Nevada  City 

37 

No.  Bloomfleld 

41.6^ 

38 

Cisco 

39.8° 

39 

Summit 

34.4'=^ 

40 

La  Porte 

28.8° 

41 

32.2° 

\9. 

Woodleaf 

69.00 

43 

Deer  Creek... 

44 

45 

6  500 
5  500 

52.38 

90.49 
57.96 

46 

Bowman's  Dam. 

Bear  Eiver  Drainage  Basin. 


47 


Bear  Valley 4  600 


Wheatland , 


Grass  Valley. 
Gold  Run 


84 

2  090 

3  222 


1904 
1907 


1904 
1907 


1907 
1907 


4.46 
14.59 


1.09 
4.67 


11.22 
10,47 


34.26 
11.10 


6.14 
3.06 


11.79 
9.61 


27.99 
35.50 


7.22 
9.64 


26.15 
21.61 


66.71 
61.19 


14.45 
17.37 


49.16 
41.69 


1.18 


0.81 


6.19 


50.0° 


39.4° 


114  THE  FLOOD  OF  MAECH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

TABLE  1.— (Continued.) 


Rainfall  Stations. 


a 

o 

No. 

Name. 

^ 

tt) 

OJ 

P 

H 

Precipitation,  in  Inches. 


PS 


S5 


Ho 

0)  g 


American  Kiver  Drainage  Basin. 


51 

Colfax 

2  421 

1904 
1907 

3.50 
9.45 

20.10 
9.75 

30.46 
19. 4f; 

44.06 
38.66 

2.401.55 

2.85 

12.36 

48.3° 

5'i^ 

Emigrant  Gap 

5  230 

1904 
1907 

3.75 
14.35 

25.10 
14.45 

31.23 

30.2(1 

60.07 
59.00 

2.00  3.50 

4.50 

19.45 

30.0° 

53 

Georgetown 

2  650 

1904 
1907 

4.79 
8.96 

36.02 
13.50 

31.17 
39.07 

51.96 
51 .53 

3.581.08 

4.90 

19.47 

42.4° 

54 

Placerville 

2109 

1904 
1907 

2.96 
8.13 

15.59 

8.15 

13.4^ 
30.54 

33.03 

36.82 

3.03  1.06 

4.28 

14.62 

47.0° 

55 

Rocklin 

249 

1904 
1907 

1.29 
5.51 

7.94 
5.71 

3.1!r 
12. 4t 

16.41 

23.68 

1.70 

0.15 

2.20 

8.65 

51.2° 

56 

Represa 

Auburn 

305 
1360 

1904 
1907 

1904 
190? 

1.15 
6.31 

2.73 

8.35 

8.33 
5.31 

13.34 

9.70 

8.55 
12.39 

11.8.- 
16.66 

18.03 
34.01 

27.90 
31.71 

57 

3.08 

0.39 

3.11 

11.23 

47.6° 

58 

Blue  Caayon 

4  395 

1904 
1907 

4.81 
13.18 

30.61 
17,95 

26.14 
35.1] 

61.56 
66.34 

4.18  4.35 

6.45 

37.33 

36.6° 

59 

Iowa  Hill 

3  825 

1904 
1907 

4.58 
11.53 

20.20 
10.13 

16.97 
24.36 

41.75 
46.01 

3.43  2.88 

3.21 

16.35 

42.6° 

60 

New  Castle 

970 

1904 
1907 

1.93 
7.09 

10.79 
6.73 

11.61 
14.10 

24.33 
27.91 

1.35  1.18 

1.78 

8.49 

50.0° 

61 

Eolsom 

252 

1904 
1907 

1.12 
5.25 

7.19 
5.65 

7.70 
11.06 

16.01 
21.96 

1.42  0.10 

2.08 

7.33 

51.0° 

a?. 

Pilot  Creek 

4  000 
3  704 

1904 
1907 

ir.04 

1907 

5.48 
14.40 

3.84 
9.45 

29.88 
11.79 

35.50 
13.24 

25.45 

32.88 

23.29 
24.05 

60.81 
59.07 

52.63 
45.74 

Towle 

63 

3.69 

15.83 

2.402.43 

i 

37.7° 

Stony  Creek  Drainage  Basin. 


65 


Fouts  Springs 

1650 

750 
254 

1904 
1907 

1904 

1904 
1907 

2.34 

14.8;. 

0.75 

9.44 
4.60 

4.79 

4.06 

12.73 
15.63 

6.22 

6.36 
3.97 

24.51 
35.08 

11.76 

Julian 

Orland 

0.57 

0.15 

0.12 



2. Hi  48.8° 

1 

Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  115 

TABLE  1.— (Continued.) 


Kainfall  Stations. 

Precipitation, 

IN  Inches. 

g 

aj 

J3 

j=     ja 

ja 

2*- 

No 

Name. 

1 

a 

■^          b. 

o  o 

t-  a 

M 

00 

03 

tJ'-O 

-^  o 
a*- 

o 

1-5 

fc         g 

^ 

s 

^5 

^2 

Cache  Creek  Dkainage  Basin. 


Bartlett  Springs 2  375 

KonoToyee 1350 

Lake  Port 1325 

Upper  Lake 1  350 

Guinda 350 

Woodland 


1904 

2.48 

19.96 

16.75 

39.19 

1904 

1.64 

8.78 

7.62 

:8.04 

1904 

1.65 

13.37 

12.72 

27.74 

1904 
1907 

1.62 
5.30 

11.19 

4.60 

10.14 
10.63 

2i.95 
20.53 

2.40 

1.73 

0.70 

7.98 

1904 
1907 

0.75 
9.30 

6.80 
1.30 

7.55 

8.84 

15.10 
19.44 

1.30 

1.00 

1.70 

7.20 

1904 
1907 

0.69 
4.45 

4.60 
3.24 

7.15 
5.90 

12.44 
13.. 59 

4.10 

i 

47.6° 
47. 9» 
50.50 


PuTA  Creek  Drainage  Basin. 


73 

Middletown 

1  300 
51 

1904 

1904 
1907 

2.52 

0.53 
4.81 

16.99 

5.05 
2.28 

27.57 

7.57 
6.69 

47.08 

13.15 

13.78 

1.25 

0.07 

2.00 

5.24 

74 

Davisville 

56.  lo 

75 

North  Lake  Port 

1450 
363 

1907 

1904 
1907 

5.45 

2.65 
10.89 

4.30 

16.08 
7.95 

12.35 

16.10 
19.50 

22.10 

34.83 
38.34 

76 

Calistoga 

0.00 

5.80 

3.85 

16.60 

52. 6» 

77 

Helen  Mine.  

2  750 

1904 
1907 

4.52 
27.21 

34.22 
11.66 

31.48 
36.73 

70.22 
75.60 

7.40 

6.64 

5.10 

28.90 

43.4° 

78 

VacavUle 

175 

1904 
1907 

1.67 
6.54 

8.61 
3.08 

11.73 

8.48 

18.10 

0.13 

2.02 

0.29 

4.81 

49. 7° 

79 

Mt.  St.  Helena  . .  .• 

2  300 

1904 
1M07 

3.37 
19.95 

28.34 
12.18 

26.14 
24.20 

57.85 
56.33 

116  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

TABLE  l.—iContinued.) 


No. 


Rainfall  Stations. 


Name. 


Precipitation, 

IN  Inches. 

,to 

si 

,q 

^' 

.a 

— ^a 

■^ 

.  +5 

n 

c3 
►-5 

.o 

b 

nt 

00 

^^ 

fe 

S 

^n 

s 

^5 

05 

S 

a 

a 

Acs 

as 


J^  :g^ 


San  Joaquin  Drainage  Basin. 


80 

Farmlngtoii 

Fresno 

Ill 
293 

1904 
1907 

1904 
1907 

0.54 
4.70 

0.57 
3.35 

4.71 
2.65 

2.49 
0.84 

4.10 
5.47 

2.f5 
1.74 

9.35 
12.82 

5.81 
6.03 

52.9° 

81 

0.12 

0.00 

0.32 

0.56 

52. 8» 

H?, 

255 

1904 
1907 

1.15 
4.40 

2.60 
0.77 

2.90 
5.56 

6.65 
10.73 

0.00 

0.45 

0.00 

2.96 

47.3° 

RS 

Las  Banas             

121 

1904 
1907 

0.25 
8.17 

1.23 

1.17 

1.28 
4.39 

2.76 
8.73 

0.00 

0.45 

0.67 

2.78 

53.0° 

84 

Mendota 

177 

1904 
1907 

0.20 
2.83 

1.70 
1.31 

1.26 
1.79 

3.16 
5.93 

0.00 

0.03 

0.00 

0.64 

55.5° 

85 

Merced  

173 

1904 
1907 

0.55 
4.25 

2.30 
3.16 

2.34 
3.68 

5.19 
11.00 

0.00 

0.00 

0.22 

2.73 

51.7' 

86 

Newman 

91 

1904 
1907 

0.23 
3.35 

1.51 
1.49 

2.33 
3.82 

4.07 
8.66 

0.15 

0.00 

0.67 

2.42 

51.8° 

87 

Stockton         

23 

1904 
1907 

0.54 
3.94 

4.09 
2.52 

3.67 
6.03 

8.30 
12.49 

0.63 

0.06 

1.22 

4.07 

51.2° 

88 

Storey 

296 

1904 
1907 

0.69 
2.70 

2.69 

0.48 

2.47 
1.35 

5.85 
4.53 

0.00 

O.CO 

0.01 

0.47 

49.8° 

89 

Tracv 

64 

1904 
1907 

0.46 
3.22 

2.10 
1.70 

1.93 

5.04 

4.49 
9.96 

0.00 

0.00 

0.70 

2.75 

48.0° 

t)0 

Westley 

90 

1904 
1907 

0.41 
5.18 

1.53 
1.39 

3.07 
3.55 

5.01 
10.12 

0.00 

0.18 

0.48 

2.31 

55.2° 

«)1 

No.  Fork 

3000 

345 
25 

46 

1904 
1907 

1907 

1907 

1904 
1907 

10.73 
14.30 

4.24 

4.64 

4.65 
6.43 

■27;9i 

9.30 

9.76 

7.72 
11.46 

Pollasky 

Lathrop 

9.19 

4.20 

3.58 

0.42 
3.23 

4.42 

0.86 

1.54 

2.65 
1.80 

W 

93 

41.5° 

94 

Antioch 

0.65 

0.03 

0.83 

4.38 

.54.8° 

Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  117 

TABLE  1.— (Continued.) 


Rainfall  Stations. 

Precipitation, 

IN  Inches. 

tteS 

. 

HSS 

a 

m 

A 

^      ja 

j: 

«  t. 

—'.a 

a£ 

No 

Name. 

^ 

1 

? 

^ 
aj 

cS 

|o 

h"f 

> 

^ 

b. 

s 

Hp 

b 

03        eS 

nJi 

a 

CO 

s  1  s  1  a 

CosuMNEs  Drainage  Basin. 


95  Shingle  Springs 1  427 


1904 
1907 


2.80   15.91 
5.05     8.50 


J2.58 
16.84 


31.29 
30.39 


MOKELUMNE    RlVER   DRAINAGE    BASIN. 


9f 

Mill  Creek 

3500 

1  500 

725 
1560 

2  326 

49 

287 

35 

1907 

1904 
1907 

1904 
1907 

1904 
1907 

1904 
1907 

1904 
1907 

1904 
1907 

1904 
1907 

10.20 

2. OS 
6.25 

2.61 

7.47 

2.44 
7.61 

4.62 
9.38 

0.60 
4.00 

0.90 
4.87 

0.72 
3.94 

8.00 

13.79 
6.25 

13.92 
5.15 

13.35 
6.29 

IB.S*? 
6.66 

6.24 
3.29 

7.05 
3.95 

5.77 

2.82 

24.45 

9.22 
13.85 

9.50 
18.01 

9.52 
15.66 

13.4? 
19.76 

5.27 
7.59 

5.00 
10.39 

4.85 
6.76 

42.65 

25.09 
26.35 

26.03 
30.63 

25.31 
29.56 

34.47 
35.80 

12.11 
15.38 

12.95 
19.21 

11.34 

2.06 

2.36 

4.04 

15.96 

41. 5» 

97 

Kennedy  Mine 

Electra 

98 

0.26 

4.55 

0.47 

10.04 

99 

Mokelumne  Hill 

West  Point 

53.2° 

48.0° 

100 

Gait    

101 

1.00 
1.10 

0.04 
0.11 
0.21 

1.47 
2.40 
1.18 

5.92 
6.27 
4.26 

102 
1(« 

Tone 

Lodi 

51.3' 
44.9» 

13.52 

0.42 

50. 70 

Calaveras  Drainage  Basin. 


101 

Milton        

660 
673 
300 

1904 
1907 

1904 
1907 

1907 

0.93 
4.76 

1.42 
5.51 

6.78 
2.53 

10.56 
4.31 

2.61 

5.30 
9.27 

7.81 
11.12 

9.38 

13.01j 
16.560.34 

19.79; 
20.94  0.92 

0.35 
0.25 

1.53 
2.62 

4.90 
6.66 

105 

Valley  Springs 

51.2° 

106 

54.2° 

118  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFOEXIA  RIVERS    [Papers. 

TABLE  1.— (Continued.) 


Rainfall  Stations. 

Precipitation,  in  Inches. 

No. 

Name. 

1 

> 

a) 

. 

P 

a 

w 
<; 

o  o 

CO 

!> 

s 
g 

1 

si 

.a 

^  1. 

■"  o 

Stanislaus  River  Drainage 

]Jasin. 

107 

Oakdale  

156 

760 
7  500 

1904 
1907 

1907 

1907 

0.70 
3.72 

5.03 

11.38 

5.00 
2.36 

6.49 

2.63 

3.44 
6.37 

17.48 

29.43 

9.14 
12.45 

29.00 

43.44 

0.59 

0.00 

1.33 

3.36 

108 

Melones 

Relief  Creek 

49.1° 

109 

Tuolumne  River  Drainage 

Basin. 

110 

111 

112 

113 
114 
115 

llli 

Jamestown 

1  471  1904 
•  1907 

3  100 '  1907 

1 

4  452  1904 

|1907 

1 

3  100  1907 

i 

850  1907 

91)  1904 

1907 

1  900  1904 

1.96 

7.82 

-  9.89 

1.87 
13.49 

9.66 

5.76 

0.33 
4.11 

1.79 
7.38 

12.96 
5.59 

4.92 

17.10 
5.82 

3.70 

4.54 

1.67 
3.00 

13.82 
5.40 

8.18 
17.27 

30.79 

19.56 
27.41 

15.95 

13.38 

3.15 
4.70 

8.63 
19.09 

23.10 
30.58 

35.60 

38.53 
46.73 

29.31 

23.68 

4.15 
11.81 

24.24 
31.87 

1.04 

1.43 

1.83 

10.68 

Tuol  Camp 

49.0° 

Crocker's  (Sequoia  P.  0.). 

Modesto 

0.34 
1.41 

1.05 
1.46 

0.76 
1.93 

3.64 
12.17 

57.3° 

1907 

47.4° 

Merced  River  Drainage  Basin. 


117 
118 

Summerdale 

Yosemite 

5  270 

3  945 

375 
126 

1 
1904      2.60;  14.96    17.09 
1907:   14.95     6.81:  37.06 

1904     2.99    13.95    12.53 
1907    11.96     3.72'  21. 9S 

1 
1907|    4.04     2.08     5.85 

1 
1904'     0  57     2.19i     3.17 

34.66 

48.82 

29.47 
S6.66 

11  97 

2.r4 

0.71 

1.90 
1.85 

13.90 

35.2° 

119 

2.61 

3.02 

13.15j  38.4° 

1-'0 

Elm  wood        . .        

4.93 

0.22 

1.64 

1907'i     3.60     0.60     3.26 

7.46 

0.00 

0.00 

53.6° 

Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  119 


TABLE  2. — Rainfall  Stations,  x\ccording  to  Elevation, 
FOR  March,  1907. 


ELE.VATION,  IN  FeET. 

0-500 

46 

Stations. 

! 
500-1  000    1  000-2  000  3  000-3  000 

10                 13                 16 
Stations.  ,  Stations.  Stations. 

3  000-4  000 

15 
Stations. 

4  000-5  000 

7 
Stations. 

5  000- 

6  Sta- 
tions. 

Basin. 

SB  ig 

t^             IS 
ft            - 

6.20  50.3° 
9.22  51.9° 

a 
o 

ft 

3 

S  a 

a 

a 
o 

If 
ft 

19.46 

^              Mean 

?       temperature. 

to 

^               Total 

~       precipitation. 

3 

it 
s  ft 

E 

a 
o 

ft 

26.42 
32.68 
34.77 
21.61 
28.46 

3 

II 

S 

<u 

37.6° 
36.9° 
39.8° 
39.4° 
37.7° 

d 
.2 

J. 

t 

il 
IE 

B 

1 

E 
ft 

5.98^49.6° 

44.6° 
43.4° 
41.6° 

4516° 

Keatner  

35.38 

35.70 
26  15 

36.03 
42.62 
35.50 
35.11 

32!  2° 

Yuba 

i9.3i' 

19  43 

28.01 

Bear 

9  t;4  50  6° 

American 

11.97  51.1° 

14.16  50.6°  16-66 

47.6° 

33.36 

36.6° 

30.20 

16.84 
14  76 

Mokelunine 

8.25  49.0° 
9  38    ... 

18  0rP3.2° 

48.0° 

19.76 

24.45 

41.5° 

Calaveras 

10  20.52  7° 

Stanislaus 

6  37  49  1°  17  48 



29.43 

Tuolumne 

4.70  57.:^° 
4.5«  53  (5° 

13  38 

18.18 

48.3° 

30.79 

...    . 

15.95 

20.98 

38 '.4° 

27.41 
27.06 

.35!  2° 

Merced 

San  Joaquin 

4.12  51.2° 

3.97  48.8° 

7.37  49.2° 

11.56|52.8° 

14.30 

Stony  Cr. 

6.22 

15.63 
10.63 
12.35 

Cache  Cr...   . 

47.6° 

PutaCr 

30.46 

43.4° 

28.66 

38.3° 

34.25 

34.7° 

Average    of    all       \ 
Stations ') 

6.65  51.0°  13.09  51.6° 

i 

16.36 

48.2° 

25.34 

43.9° 

28.61 

A  comparison  of  the  precipitation  from  January  to  March,  1907, 
with  that  for  the  same  period  and  stations  in  1904,  shows  that,  for  the 
average  of  all  stations  in  the  water-shed,  the  precipitation  was  greater 
in  1907  than  in  1904,  and  that  the  difference  increases  from  the  north 
toward  the  south,  but  the  percentage  in  favor  of  the  former  is  quite 
small.  An  examination  of  Table  1  shows  that  the  precipitation  for 
January,  1904,  v;as  quite  light  compared  with  that  of  January,  1907, 
while  the  precipitation  for  February,  1904,  was  much  heavier  than 
for  February,  1907.  The  comparison  for  March,  however,  is  of  most 
importance,  as  regards  the  floods  of  1904  and  1907.  Such  a  com- 
parison is  made  in  Table  4,  where  it  is  seen  that,  with  the  exception 
of  the  Sacramento  River  Basin,  the  precipitation  throughout  the  water- 
shed was  much  greater  in  1907.  In  the  basins  of  the  tributaries  of  the 
Sacramento  River  from  the  east,  the  rainfall  in  March,  1907,  was 
from  20  to  41%  greater  than  in  March,  1904,  while  for  basins  on  the 
vrest  it  is  only  from  2  to  3%  greater.  For  the  San  Joaquin  River 
and  its  tributaries  tjie  percentage  is  much  greater,  ranging  from  about 
50  to  80  per  cent.     The  distribution  of  the  precipitation  during  the 


120 


THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 


TABLE  o. — Resitlts  at  Twenty-four  Stations.  Ranging  rx 
Altitude  fkom  60  to  5  270  Feet. 


Drainage. 

I 

-i 

o 
o 

Precipitation 

IN  March,  1907. 

Temperature. 

Month. 

March 

17th-26th. 

March 
17th-i9th. 

Month. 

March 
17th-20th. 

station. 

** 

c 

0/ 

s 

o 

up 

W)    . 

be   . 

tvH 

S  u 

a 
a 

go 

5 

a 

a 

t 

^  o 

P^ 

s:- 

g. 

hJ 

H 

c3 

CL,  O 

Alturas 

Pit 

3 

4.13 

59 

35 

35.6^ 

43. 0» 

13° 

Auburn 

American. . . 

36 

16.66 

224 

67 

33 

107 

47.6° 

45.8° 

19° 

Cedai-ville 

Pit 

13 

3.31 

124 

48 

16 

37 

34.4° 

42.0° 

19° 

Chico 

Sacramento. 
Sacramento. 

37 
24 

8.03 
3.80 

197 
28 

60 

58 

32 
21 

95 
27 

49.2° 
49.9° 

53.1° 

57.2° 

16° 

Colusa.. 

18° 

Fresno 

San  Joaquin 

20 

1.74 

32 

32 

25 

33 

52.8= 

58.1° 

18° 

Greenville 

Feather 

13 

24.51 

371 

81 

54 

254 

37.8° 

40.5° 

14° 

Merced 

San  Joaquin 

33 

3.68 

1.56 

74 

6 

15 

51.7° 

52.5° 

25° 

Milton 

Calaveras. . . 
Yuba 

17 
15 

9.27 
24.62 

133 
193 

53 
69 

24 

38 

56 
111 

51.2° 
41.6° 

58.3° 

48.2° 

13° 

Nevada  City. . 

19° 

No.  BlmHd.... 

Yuba 

1(1 

28.64 

254 

74 

44 

1.56 

39.8° 

45.5° 

23° 

Palermo 

Feather 

IB 

8.80 

247 

50 

23 

80 

50.8° 

55.5° 

17° 

Quincy 

Red  Bluff 

Feather 

la 

30.15 

3.53 

85 

54 

244 

35.8° 

41.4° 

16° 

Sacramento. 

30 

5.92 

81 

50 

15 

27 

48.4° 

52.2° 

10° 

Redding 

Sacranifiito. 

3^ 

7.28 

53 

63 

32 

49 

49.0° 

52.9° 

10' 

Sacramento... 

Sacraiiit'iito. 

30 

7.28 

148 

65 

37 

92 

50.9° 

56.6°        12° 

Shasta 

Sacramento. 

11 

14.47 

167 

76 

33 

88 

48.6° 

47.4° 

19° 

Stockton 

San  Joaquin 

36 

6.03 

164 

67 

30 

79 

51.2° 

.57.4° 

18° 

Summerdale  . 

Merced 

11 

27.06 

194 

51 

17 

50 

35.2° 

39.4° 

9° 

Upper  Lake.. 
Wheatland. . . . 

Cache 

as 

10.63 

238 

75 

45 

152 

49.6° 

50.8° 

16° 

Bear 

ao 

9.64 

247 

64 

33 

109 

50.6° 

56.7° 

17° 

Willows 

Sacramento. 

28 

3.63 

119 

55 

24 

53 

49.1° 

54.7° 

17° 

Yosemite 

Merced 

3 

20.98 

63 

31 

38.4° 

39.5° 

17° 

Georgetown.. 

American  . . . 

34 

29.07 

209 

07 

33 

102 

42.4= 

Average  of    a 

bove  24  Sta- 

tions 

! 

21 

12.43 

179 

63 

31 

87 

45.5° 

50  7°  ;     16° 

Average   of  7 

Stations  in 

Basin 

12.96 

185 

46.5° 

1 

Note  :— These  selected  stations  are  fairly  representative  as  regards  both  temperature 
and  precipitation.  It  is  observed  that  the  average  of  the  mean  temperature,  March  17th- 
20th,  is  5.2°  above  that  for  the  month. 

month,  however,  is  of  most  vital  significance.  For  March,  1904,  the 
precipitation  is  distributed  quite  evenly  throughout  the  entire  month, 
though  the  intensity  is  noticeably  greater  during  the  equinoctial  week. 
For  March,  1907,  however,  not  only  is  the  total  precipitation  for  the 
month  considerably  greater  than  in  1904,  but  its  periodic  occurrence 
in  a  series  of  storms  is  more  pronounced.  During  the  10-day  period 
centering  about  the  equinox,  the  intensity  was  so  great  that  about  70% 
of  the  total  precipitation  for  the  month  occurred  in  this  time,  while 
more  than  30%  of  it  was  recorded  on  March  17th,  18th,  and  19th. 


Papers.]    THE  FLOOD  OF  MAECH,  1907,  IN  CALIFORNIA  RIVERS  121 


TABLE  4. — Comparison   of  Precipitation   in   Sacramento  and   San 
Joaquin  Basins,  for  March,  1904,  and  March,  1907. 


River  basin. 

Number  of 

precipitation 

stations. 

Precipitation. 

Percentage  op 
Difference. 

1904. 

1907. 

1904. 

1907. 

Sacramento 

17 

10 
8 
2 

11 
2 
3 
5 

13 
1 
7 
2 
1 
4 
3 

ii.a* 

16.52 

22.71 

17.60 

20.46 

9.54 

8.28 

18.60 

3.34 

12.58 

6.56 

8.12 

3.44 

9.63 

10.60 

10.13 
23.24 
27.20 
22.57 
25.67 

9.80 

8.46 
19.12 

4.86 
16.84 
10.20 
13.15 

6.37 
17.12 
17.10 

12 

Feather 

41 

Yuba    

20 

Bear 

28 

American 

25 

Stony  

Cache 

3 
2 

Puta 

3 

San  Joaquin 

45 

Cosumues  

34 

Mokelumne 

56 

Calaveras 

Stanislaus 

62 

85 

Tuolumne 

78 

flierced 

61 

Flood  Flow  of  Streams. 

On  the  following  pages  is  recorded  the  daily  flow  of  the  various 
streams  in  the  Sacramento  and  San  Joaquin  Basins  for  the  11  days, 
March  16th-26th,  also  the  mean  daily  flow  for  the  4-day  period,  March 
18th-21st,  at  all  places  where  gauging  stations  were  maintained.  The 
figures  given  herein  are  not  the  final  figures  as  they  may  appear  in 
the  Annual  Report  of  the  United  States  Geological  Survey,  but  they 
will  not  differ  materially  from  them.  In  all  cases  it  is  believed  that 
the  estimates  are  quite  conservative,  and  rather  inclined  to  be  too  low 
than  too  high. 

Pit  River. — This  river  drains  a  long,  comparatively  narrow,  and 
high  mountainous  area  in  the  northeastern  part  of  the  Sacramento 
Basin.  In  this  area  are  several  large  reservoir  sites.  Those  surveyed, 
to  June,  1905,  have  a  capacity  of  6  000  000  acre-ft.,  but  Big  Valley 
Eeservoir,  above  Bieber,  with  a  capacity  of  3  196  000  acre-ft.,  is  in  all 
probability  the  only  one  that  could  be  utilized  for  flood  control.  The 
precipitation  in  this  basin  above  Bieber  was  comparatively  light,  and 
occurred  mainly  as  snow,  so  that  the  run-off  per  square  mile  was  small. 

The  gauging  station  is  about  12  miles  below  Bieber  and  about  70 
miles  in  a  direct  line  above  the  mouth  of  the  McCloud  River.  The 
area  above  this  gauging  station  is  2  950  sq.  miles.  Table  5  contains 
data  on  the  flood  flow  at  this  station  during  the  flood  of  1907. 


122  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

TABLE  5. — Flow  of  Pit  Kivee,  near  Bieber. 


Date,  1907. 

Gauge  height. 

Di.scharge,  in  cubic  feet 
per  second. 

March  16th 

6.5 
12.8 
15.5 
16.4 
15.5* 
14.0* 
11.5* 
9.7 
8.5 
7.4 
7.3* 

2  770 

"     irth 

17  400 

'•     18th 

•'     19th 

25  000 
27  500 

'•     20th 

25  000 

■'     21.st 

20  800 

•'     32d 

13  800 

'•    2:M 

S810 

"    24th 

6  110 

'•     25th 

4  160 

"     26th 

4  000 

Period. 

Mean  daily  discharge,  in 
cubic  feet  per  second. 

Total  run-oft",  in  acre-feet. 

February 

4  180 

6  940 

24  600 

232  000 

Blarch 

427  000 

Blarch  18th-Slst 

195  000 

*  Estimated. 

The  mean  rate  of  flow  for  the  24  hours  when  it  was  greatest  was 
9.3  cu.  ft.  per  sec.  per  sq.  mile,  and  the  mean  rate  for  the  4  consecutive 
days,  March  18th-21st,  was  8.3  cu.  ft.  per  sec.  per  sq.  mile.  This  small 
run-off  was  due  to  light  precipitation  and  to  the  slow  melting  of  the 
snow. 

McGloud  River. — McCloud  River,  the  principal  tributary  of  the 
Pit  River,  drains  a  long,  narrow,  mountainous,  timbered  strip  of  about 
676  sq.  miles  on  the  north  side  of  the  Pit  River  Basin,  including  the 
southern  and  eastern  slopes  of  Mount  Shasta.  Its  low-water  flow  is 
remarkably  large,  never  having  been  less  than  1  200  cu.  ft.  per  sec.  at 
the  gauging  station  in  4  years. 

The  gauging  station  is  14  miles  east  of  Baird  Spur,  on  the  South- 
ern Pacific  Railroad,  at  Gregory  Post-Office,  and  the  drainage  area 
above  it  is  608  sq.  miles.  Table  6  contains  data  on  the  flow  at  this 
station  during  the  flood  of  1907. 

The  mean  rate  of  flow  at  this  station  for  the  24  hours  w^hen  it  was 
greatest  was  50.0  cu.  ft.  per  sec.  per  sq.  mile,  and  the  mean  for  the 
4  consecutive  days,  March  18th-21st,  was  35.5  cu.  ft.  per  sec.  per  sq. 
uii'o. 

Upper  Sacramento  River. — The  gauging  station  on  the  Upper 
Sacramento  is  in  the  foot-hills  near  Iron  Canyon,  4  miles  above  Red 
Bluff,  at  an  elevation  of  about  310  ft.  above  sea  level.     The  drainage 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS 


123 


area  above  it  includes  9  300  sq.  miles  of  mountains  and  foot-hills. 
About  38%  of  this  area  is  above  the  stations  on  Pit  and  McCloud 
Rivers.  Table  7  contains  data  on  the  flow  of  the  river  at  this  place 
during  the  floods  of  1907  and  1904.    , 

TABLE  (). — Flow  of  McOloud  River,  near  Gregory. 


Date,  1907. 

Gauge  height. 

March    16th 

l~th 

18th 

19th 

'30th 

'^.8 
4.0 
9.4 
13.0 
10.65 

2l!-t 

22d . . 
23d.. 
24th. 
25th. 
26th. 


.5 
5.9 
5.5 
4.9 
4.55 
3.95 


Discharge,  in  cubic  feet 
per  second. 


2  460 

4  310 
19  200 
30  too 
24  200 
12  700 

8  360 
7  400 
6  060 

5  300 
4  120 


Period. 

Mean  daily  discharge,  in  cubic 
feet  per  second. 

Total  run  off,  in 
acre-feet. 

February.  .        

5  490 

305  009 

March...     .            

5  990 
21  600 

1 

368  000 

March  18th-21st 

171  000 

TABLE  7. — Flow  of  Upper  Sacramento  River, 

NEAR  Red  Bluff. 

Date.  1907. 


March  16th 

"  17th 

"  18th 

•'  19th 

•'  20th 

••  21st. 

■'  22d.. 

•'  23d., 

"  24th 

^'  25th 

''  26th 


Gauge 
height. 


10.0 

31.4 

36.05 

28.7* 

22.85 

18.4 

31.65 

16.8 

14.3 

13.25 


Discharge,  in 
cubic  feet 
per  second. 


23  600 

.39  100 

118  000 

164  000 

192  000 

132  000 

92  900 

120  000 

80  800 

64  000 

57  100  . 


Date,  1904. 


Gauge 
height. 


Mar 


eh  7th 16.30 

8th I  24.40t 

9th !  18.95 

10th :  17.90 

11th ;  15.80 

12th I  14.70 

13lh 13.30 

14th 1  15.80 

15th I  17.25 

l'6th 18.30 


Discharge,  in 

cubic  feet 
.  per  second. 


77  200 
147  180 
97  300 
88  940 
73  700 
66  220 
57  400 
73  700 
84  050 
92  040 


Period. 

Mean  daily  discharge,  in 
cubic  feet  per  second. 

Total  run-off,  in 
acre-feet. 

February,  19C4 

46  300 
73  300 
45  700 
55  700 
152  000 

2  670  000 

March,  1904 

February.  1907 

March.  1907 

March  18th-21st,  1907 

4  510  000 

2  540  000 

3  430  000 
1  200  000 

*  Maximum  stage,  29.4  ft.;  discharge.  204  000  cu.  ft.  per  sec.  at  2  p.  m. 
t  February   16th,  the  stage  was  38.00  ft.:  maximum  stage.  31. (i  tt.:   dischan 
cu.  ft.  per  sec.  in  the  evening.    On  the  15th  the  stage  was  17.4  ft.  and  on  the  17th,  1 


■ge,  234  OCO 
2  ft. 


124  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 


The  mean  rate  of  flow  at  this  station,  for  the  24  hours  when  it  was 
greatest,  was  20.T  cu.  ft.  per  sec.  per  sq.  mile,  and  the  mean  for  the 
4  days,  March  18th-21st,  was  16.3  cu.  ft.  per  sec.  per  sq.  mile. 

Attention  is  directed  to  the  fact,  shown  in  Table  7,  that,  although 
the  flow  for  4  days  of  the  1907  flood  was  greater  than  for  any  4  con- 
secutive days  of  1904,  the  total  flow  for  March,  1907,  is  only  76%  of 
that  for  March,  1904. 

By  comparing  the  discharges  and  drainage  areas  above  the  gauging 
stations  on  the  Pit,  McCloud,  and  Sacramento  Rivers,  it  is  seen  that, 
although  the  drainage  area  above  the  stations  on  the  Pit  and  McCloud 
Rivers  is  38%  of  that  above  the  station  on  the  Sacramento,  the  com- 
bined flow  at  these  two  stations  is  only  21%  of  that  of  the  Sacramento 
during  February,  and  23%  during  March.  This  condition  is  due 
largely  to  the  slower  melting  of  snow  during  these  months  in  the 
higher  parts  of  the  basin. 

Feather  River. — The  Feather,  the  largest  tributary  of  the  Sacra- 
mento, derives  its  water  from  melting  snow  in  the  high  Sierras,  the 
highest  point  in  its  basin  being  more  than  10  000  ft.  above  sea  level. 
The  main  river  is  formed  by  the  union  of  three  streams,  the  North, 
Middle,  and  South  Forks,  above  Oroville.  Its  principal  tributaries 
are  the  Yuba  and  Bear  Rivers,  which  enter  it  below  Oroville. 

TABLE  8. — Flow  of  Feather  Kiver,  at  Oeoville. 


Period. 

Mean  daily  discharge,  in 
cubic  feet  per  second 

Total  nan-off,  in 
acre  feet. 

February,  1904 

27  800 
39  500 
21500 
36  000 
97  300 

1  600  000 

March,  1904 

2  430  000 

February,  1907 

1  190  000 

March,  1907 

2  210  000 

March  18th-21st,  1907 

770  000 

*  Estimated.       t  Maximum,  about  1  a.  m.,  185000  cu.  ft.  per  sec. 


Papers.]    the  FLOOD  OF  MARCH,  1907,  IX  CALIFORNIA  RIVERS 


125 


The  ganging  station  is  on  the  main  stream,  in  the  foot-hills  at 
Oroville.  The  drainage  area  above  it  is  3  640  sq.  miles.  Table  8  con- 
tains data  on  the  flood  flow  at  this  station  during  the  floods  of  1907 
and  1904. 

The  mean  rate  of  flow  for  the  24  hours  when  it  was  greatest  was 
35.6  cu.  ft.  per  sec.  per  sq.  mile,  and  the  mean  for  the  4  days,  March 
18th-21st,  was  26.7  cu.  ft.  per  sec.  per  sq.  mile. 

It  is  seen  that,  although  the  maximum  daily  discharge  in  1904  is 
only  73%  of  that  in  1907,  the  total  flow  for  February  and  March  is 
greater  in  1904  than  in  1907  by  29%  and  10%,  respectively. 

The  other  four  largest  floods  in  the  stream  on  record,  or  even  re- 
called by  the  oldest  inhabitants  living  along  it,  occurred  in  1849, 
!1853,  1861,  and  1881.  In  none  of  these  floods,  however,  was  the  water 
as  high  at  Oroville  as  in  March,  1907.  This  may  have  been  due  in 
part  or  entirely  to  the  filling  of  the  river  channel  at  and  below  Oro- 
ville with  mining  debris.  About  one-half  of  the  Town  of  Oroville  was 
flooded  for  3  days.  The  water  was  about  3  ft.  deep  on  the  floor  of  the 
Union  Hotel.  The  highway  bridge  and  the  Northern  Electric  Rail- 
way bridge  in  Oroville  were  swept  away,  and  also  other  bridges  along 
this  stream. 

The  great  range  of  river  stage  and  its  rapid  fluctuations  are  shown 
by  Table  9,  the  gauge  record  at  Big  Bend,  15  miles  above  Oroville. 


TABLE  9. — GrAUGE  Eecord  on  Feather  Eiver,  at  Big  Bend.* 


Day. 

Hour. 

Gauge  height. 

March  14th 

7  a.m. 

8  " 

8  " 

9  '■ 

1  P.  M. 
6    '• 

3  A.  M. 

10    " 

4  P.  M. 

1  A.  M. 
10     " 

10     " 
10    '• 

5  P.  M. 

10  A.  M. 

6  9 

15th 

6  6 

16th 

6  6 

17th 

10  0 

"      17th 

13  0 

171h 

20  0 

18th 

25  0 

iwth 

31  0 

18th 

32.5 

"       19th 

36  0 

"      19th 

34  5 

"      20th 

28  0 

"      21st  

23  0 

"      asd 

18  0 

'■      25th 

10  0 

Note:  Low-water  reading,  2  ft. 

*  Data  furnished  by  Great  Western  Power  Company,  through  Mr.  L.  J.  Bevan. 

Indian  Creeh. — Indian  Creek  is  a  tributary  of  the  North  Fork  of 
the  Feather  River,  and  its  water-shed  is  at  a  high  altitude.    The  gaug- 


126 


THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 


ing  station  is  about  1^  miles  below  the  Town  of  Crescent  Mills.  The 
drainage  area  above  this  station  is  740  sq.  miles,  the  larger  part  of  it 
being  at  an  elevation  of  more  than  5  000  ft. 

TABLE  Xo.  IO.^Flow  of  Indian  Ckeek,  near  Ceescent  Mills. 


Date,  1907. 

Gauge  height. 

Discharge,  in  cubic  feet 
per  second. 

March  16th 

17th 

4.4 
7.1 
17.0 

19.7* 
17.9 
14.7 
10.95 
9.0 
7.8 
7.7 
7.5 

905 
2  570 

"       I8th 

9  500 

19th 

11  500 

'•      20th 

10100 

"       21st 

7  890 

^'       22d 

5  365 

23d.. 

3  900 

"       24th 

3  060 

"       25th 

2  990 

"       26th 

2  850 

Period. 

February 

March 
March'l8th-2ist.!!! 


Mean  daily  discharge,  in 
cubic  feet  per  second. 


Total  ruu-off .  in  acre-feet. 


2  210 
2  940 

9  750 


123  000 
IM  000 

77  400 


*  Maximum,  20.2  ft. 


The  mean  rate  of  flow  for  the  24  hours  when  it  was  greatest  was 
15.5  cu.  ft.  per  sec.  per  sq.  mile. 

It  is  seen  that  the  maximum  run-off  per  square  mile  during  this 
flood  is  less  than  one-half  of  that  from  the  water-shed  of  the  Feather 
River  above  Oroville,  due  to  the  slower  melting  of  the  snow  at  high 
altitudes. 

Yuha  River. — The  Yuba  is  the  largest  tributary  of  the  Feather 
River,  entering  it  at  Marysville,  30  miles  above  the  junction  of  the 
Feather  and  Sacramento  Rivers  and  26  miles  below  Oroville.  The  en- 
tire area  drained  by  it  is  about  1  330  sq.  miles,  of  which  1  220  sq.  miles 
are  above  the  gauging  station  near  Smartsville.  The  basin  is  com- 
paratively long  and  narrow,  the  highest  point  having  an  elevation  of 
9  000  ft.,  which  is  not  as  great  as  that  of  the  Feather  River,  the  high- 
est point  of  which  is  more  than  10  000  ft.  A  large  part  of  the  basin 
is  more  than  5  000  ft.  above  sea  level.  Table  11  contains  data  on  the 
flow  of  this  stream  at  Smartsvillo  during  the  floods  of  1907  and  1904. 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  127 

Table  11.— Flow  of  Yuba  River,  at  Smartsville. 


Discharge,  in 

Discharge,  in 
cubic  feet 

Date,  1907. 

Gauge  height. 

cubicj  feet                Date,  1904. 

per  second. 

per  second. 

Mrtrnh  Kith !. 

14.3 
24.0 

6  600            Febr 
56  000 

uary  16th 

17th 

58  000 

'     17th 

41  000 

'     18th 

27.9 

85  000 

18th 

17  880 

•     19th 

29.?* 

100  000 

19th 

13  340 

^     20th 

24.0 

60  000 

20th 

9  350 

'     21st 

18.5 

27  000 

21st 

9  350 

•     22d 

15.9 

14  000 

22d 

59  800 

•     23d 

16.4 

16  500 

23d 

27  660 

■     S4th 

15.0 

11  OCO 

24th 

59  800 

*■     25th 

14.5 

9  900           i 

25th 

24  080 

"     26th 

14.1 

8  900 

Period. 


Mean  daily  discharge,  in 
cubic  feet  per  second. 


Total  runoff,  in  acre-feet. 


February.  1904 

March,  1904 

February.  1907 

March,  1907 

March  18th-21st,  1907, 


14  900 

15  400 
14  100 
17  300 
68  000 


858  000 
947  000 
783  000 
1060  000 
537  000 


*  Maximum  stage,  29.5  ft.  about  2  p.  m. 

The  mean  rate  of  flow  for  the  24  hours  when  it  was  greatest  was 
82.0  cu.  ft.  per  sec.  per  sq.  mile.  The  maximum  daily  discharge  of 
this  stream  was  67%  greater  in  1907  than  in  1904.  The  total  dis- 
charge for  March,  1907,  is  larger  than  for  March,  1904,  but  the  total 
for  February  and  March  combined  is  about  the  same  for  the  two  years. 
The  effect  of  rapid  melting  of  snow  in  the  middle  altitudes  is  clearly 
shown  here  by  the  large  run-off  per  square  mile. 

Bear  River. — The  Bear  is  the  most  southern  tributary  of  the 
Feather  River,  entering  it  about  12  miles  above  the  mouth.  It  drains 
an  area  of  about  290  sq.  miles,  of  which  263  sq.  miles  are  above  the 
gauging  station  at  Van  Trent,  8  miles  above  Wheatland.  Its  head- 
waters do  not  reach  back  to  the  crest  of  the  Sierras,  and,  as  much  of 
its  drainage  basin  is  deforested,  it  is  more  torrential  than  the  main 
stream.  The  greatest  altitude  in  the  basin  is  about  5  500  ft.  Table 
12  contains  data  on  its  flow  during  the  1907  flood. 

The  mean  rate  of  flow  for  March  19th  is  106.5  cu.  ft.  per  sec.  per 
sq.  mile,  and  the  mean  for  March  17th-20th  is  75.3  cu.  ft.  per  sec.  per 
sq.  mile. 

It  will  be  noticed  that  the  run-off  per  square  mile  was  106.5  cu. 
ft.  per  sec.  on  March  19th,  and  102.7  cu.  ft.  per  sec.  on  February  2d. 


128  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

TABLE  12. — Flow  of  Bear  River,  at  Van  Trent. 


Date,  1907. 

Gauge  height. 

Discharge,  in  cubic  feet 
per  second. 

March    16tli 

4.9 
13.95 
12.75 
17.8 
13.6 
9.6 
8.8 
13.2 
9.7 
9.4 
8.1 

900 

17th 

18  200 

18th 

15  500 

19th 

28  000 

20th 

17  400 

'        21st  

8  400 

'        22d 

6  600 

'        23d 

16  500 

24th 

8  600 

25th 

8  000 

26th 

5000 

Period. 

Mean  daily  discharge,  in  cubic 
feet  per  second. 

Total  run-off,  in 
acre-feet. 

February 

3  460 
5  570 
19  800 

192  000 

March 

;i42  000 

March  17th-20th 

157  000 

Note  :    On  February  2d  the  discharge  was  27  000  cu.  ft.  per  see. 

American  River. — The  American  River  drains  an  area  of  about 
2  000  sq.  miles,  directly  south  of  Bear  River  Basin  and  north  of 
Cosumnes  River  Basin.  It  has  three  main  forks,  two  of  which  head 
at  an  elevation  of  about  9  000  ft.  above  sea  level,  while  the  South 
Fork  reaches  back  to  an  elevation  of  more  than  9  600  ft.  The  gaug- 
ing station  is  at  Fair  Oaks,  and  the  drainage  area  above  it  is  1910 
sq.  miles,  a  large  part  of  which  has  an  altitude  of  more  than  5  000  ft. 
Table  13  contains  data  on  the  flow  of  this  stream  during  the  flood  of 
1907. 

TABLE  13. — Elov^t  of  American  River,  at  Fair  Oaks. 


Date,  1907. 

Gauge  height. 

Discharge,  in  cubic 
feet  per  second. 

March  Ifith 

6.1 
13.40 
20.60 
27.6*+ 
23.9* 
21.0* 
18.4* 
13.5 
13.25 
12.80 
11.50 

6  800 

'       17th 

33  000 

'       18th 

63  200 

'      19th 

93  000 

'      20t.h 

'      2lst 

'      22d..... 

'      23d 

'      24th..... 

'      25th 

26th 

77  000 
65  000 
54  000 
33  400 
32  300 
38  400 
85  000 

Period. 

Mean  daily  discharge,  in 
cubic  feet  per  second 

Total  run-off,  in 
acre-feet. 

February  

14  200 
23  200 
74  600 

789  000 

March ; . . 

1  430  0(X) 

March  18th-21st 

594  000 

*  Bridges  and  gauges  washed  away.    Gauge  height  estimated. 
+  Maximum  stage,  30.2 ft., about  5  a.  m. 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS 


129 


The  greatest  mean  daily  discharge  at  this  station  was  48.7  cu.  ft. 
per  sec.  per  sq.  mile,  and  the  greatest  4-day  mean  was  39.1  cu.  ft.  per 
sec.  per  sq.  mile. 

Attention  is  called  to  the  fact  that  the  run-off  per  square  mile  for 
the  American  River  is  about  50%  greater  than  that  of  the  Feather 
River,  although  the  percentage  of  each  basin  with  an  altitude  exceed- 
ing 5  000  ft.  is  about  the  same. 

Btony,  Cache,  and  Puta  Creehs. — These  three  streams  are  the 
largest  tributaries  of  the  Sacramento  River  from  the  west.  They 
drain  the  eastern  slope  of  the  Coast  Range,  and  are  torrential.  Stony 
Creek  is  the  only  one  of  the  three  that  flows  directly  into  the  Sacra- 
mento River;  the  other  two  empty  into  Yolo  Basin.  The  gauging 
station  on  Stony  Creek  is  in  the  foot-hills,  near  Fruto,  and  the  drain- 
age area  above  it  comprises  601  sq.  miles.  The  station  on  Cache 
Creek  is  near  Yolo,  and  the  area  above  it  is  1  230  sq.  miles.  Puta 
Creek  station  is  at  Winters,  and  the  area  above  it  is  805  sq.  miles. 
Table  14  contains  data  on  the  flow  of  these  streams  during  the  flood 
of  1907. 

TABLE  14. — Flow  of  Stony,  Cache,  and  Puta  Creeks. 


Date,  1907. 

Stony  Creek. 

Cache  Creek. 

Puta  Creek. 

Gauge. 

Discharge. 

Gauge.      Discharge. 

Gauge. 

Discharge. 

March  17th 

t 

9.45               6  100 
14.25       '       25  000 
13.15       i       20  000 
11.8        !       13  450 

9.8                 6  810 

6.80                3  950 

IK  .W 

8  800 

18th 

19.45               13  500               21.60 
25.90t             19  000               23.65 
18.20               12  500               16.15 
12.65                 7  820                12.35 
12.00                7  300               11  90 
20.85               14  800       I        26  60* 
19.30               13  400       i         15.60 
16.15               10  800               14.75 
12.55        1        7  750              11.40 

19  800 
24  700 

"        19th.     . 

20th 

10  000 

21st 

5  460 

22d 

7.75 

3  350 

5  000 

"        23d 

11.55 
8.7 
8.15 
7.75 

12  300 
4  760 
3  910 
3  3.50 

31  500 

24th 

9  200 

25th 

8  100 

"        26th 

4  500 

Period. 

Mean  daily  discharge,  in                        Total  run-off,  in 
cubic  feet  per  second.                                acre-feet. 

Stony  Cr. 

Cache  Cr. 

1 
Puta  Cr.      StonyCr. 

Cache  Cr. 

Puta  Cr. 

February 

3  330 

4  450 
16  300 

3  320 

5  310 

13  200 

1  740            185  000 
5  030             273  000 
15  000       ,      129  000 

i 

129  000 
326  000 
105  000 

96  600 

March 

309  000 

March  ]8th-21st 

119  000 

*  Maximum,  28.15  ft.,  about  noon.         t  Maximum,  about  26.4  ft.,  during  night. 


IBP 


TSTF  TM»Q1>  iW  mAUCM..  l!?(<^.,  IST  CJOiEPCatXIlA  iETTiESlS   II'Apers. 


The  5jy»JKsr  flfio^j"  xh^  «f  Af*^  I^r  sgaare  amle  'was,  ico-  S*«m7 
Qyftflg,  41.6  pu-  ft.  i*er  sec.;  f«r  (Oadhe  Oreek,  laJ>  'Ca.  ift  iper  sec;  fw 
Puia  Crreei;.  BS-1  cm.  it  iper  sec  The  smaiH  teed-jsJI  l£roan  Cache  Basin 
is  ;     -  '  -        ,  ■  -    ^  ^  ~   ^-   -~   ""■'  Lake. 

_  ■  '   "     no  i-fll^ 

ine  SEfltann  imfi  Itieen  esiaihiyhefl  fln  tSm?  San  Jnagrnxii  Jtiver,  "bacaise  of 
inahiEty  *r   '  -  In  liie  fall  of  -  ■'  '  -  "^er.  a 

saodat  'wai^  .^  .    -^--.  _  __   .^..  -  ... s  near  J^oIlaskT,  ...  ...  _.  inils 

nnrrieasi  ki  TTesna.  ^«ehsre  ifair  cian^ixians  dhtadn.  The  drainage  area 
flbpve  tV>>s  scaiiun  is  1  ^€4f*  sij.  unolfe.  Itn  TTtrifciirng'  say  fSdjns^  jsf  liie  Tmn- 
ttP  inr  Marek  iStb-T"  -  ■  f  IP  <m.  ft  per  sec  par  sq.  ttiiTp  "has 

hesn  Tsed.  This  Tr--.  -  -■  .  tiie  TaTvSS  in  liie  hasins  to  xhe  nartii 
Hnfi  simiiiu  and  k  lefie^ed  ip  "he  gioiff  eanservatrve. 

~  •?  Eirer  has  a  "rery  nairo^  ami  very 

Itc.^    -.■.-.-.i.^.    -..^- ._.__.    ;^.,.:a^   easrwrard  to  -the  surmnTt  of  liie 

Siecras.  i'Tom  lie  Irw  f  fiot-MIk  "to  tSk  jBmcttaan  nS.  liie  liiree  hrandh^ 
athcmt  3C'  naD^  ulicw^.  aife  fiaBiii  k  Sl  ifarciacl  eanjnan  wiilfti  'A  mamiiiiiirna 
"wj '  "      -' '  '       ■"'■es.    A  larse  pere-:       -  :■  Trtasn  r;    -   =   '"■•^2 

7  '  '     :'.  in  eievanioiL  ei  ■     -e  naoire  i_    -  '.'0 

T  A"RTT   ~  f  — T"liOW  OF  5fE(!1EEUni5TE  ^IVI2L,  3CEAE    CliE3i03PT5.. 


lajtt.  ^iir. 


■Eiaiure  neiidit. 


per  sacfltnfi. 


SBKdi  J<n 


IS  fMTi 


Sfl. 


SOW 

112-200 
XTiOOO 
ItSOOO 

13  aw) 

*!H0O 
6300 

%4m 


reiamary 

MHwdtt. 


feei  "Der  spconcL- 


2S3t) 


Total  nm-ofi.  Sn^aert-ir-WL. 


106  <sw 


A-mwriif^ann  aood  StaaoMlaiiK  ISir^eas.  amd  emptaes  njiinip  lihe  Sana  -Joagnin 


It  1^  te 

U  !•  i*  it 
t#  !•  U 


\.     ^    - 


t  r  r  « 


TflE  1^  -WED 


ii    Jn  jifi  I'MrraB-  ingmtr^' 


snsest  :a35in-  Ti»^ 


-^irXffH.  if  taifec  "H  "figr  :agjr  "iitSC  -&&  -rrgt.'y.iriiiiri    THHITf  fi-*<'rijmj& 
"Hft-  =eE^'-'  "'  "'^^  ~"  "-^-^SHei.     ^^'-    . —    ' — 

.  -^    ^.r  "Eat  3ftr..   -  -—  —  -    ^. -    _  -  -^  IHl 


Kt  si^i^srinrt  tr 


TA'ggry  3€. — 5^F«r  ((ffl^  C>il«EM9ffiE  ^^FTi^. 


iCi  t!  Wrm^  -T     ^_ 


atefc-SisL 


.:s:aai 


3«r  S5K-  ^er  sc.  TTfTfc-  imt  'her  j!gigihHfeL  4r-£a?-  ttpstt  -fi;^  S^  ji.  ±i  T^r 

^K-  5«2r  St.  Trr<*.     TTie  Ss^tl. 

rnnt  tsta.  m.  -ttp-  m-i^-«?erT2j  :  j  :      _ 

TttranisL  ^'^tw^'  *££•  ^Biait.     JL  ^!°EznL3r  y»iiiJ'''iiJ'  4;.ij»iT"iTTr   TBt^  i^st  -sruu— 


130 


THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 


The  greatest  daily  rate  of  flow  per  square  mile  was,  for  Stony 
Creek,  41.6  cu.  ft.  per  sec;  for  Cache  Creek,  15.5  cu.  ft.  per  sec;  for 
Puta  Creek,  39.1  cu.  ft.  per  see.  The  small  run-off  from  Cache  Basin 
is  attributed  to  the  storage  and  regulation  effects  of  Clear  Lake. 

San  Joaquin  River. — Prior  to  the  flood  of  March,  1907,  no  gaug- 
ing station  had  been  established  on  the  San  Joaquin  Kiver,  because  of 
inability  to  find  a  satisfactory  section.  In  the  fall  of  1907,  however,  a 
station  was  established  in  the  foot-hills  near  Pollasky,  about  20  miles 
northeast  of  Fresno,  where  fair  conditions  obtain.  The  drainage  area 
above  this  station  is  1  640  sq.  miles.  In  making  an  estimate  of  the  run- 
off for  March  18th-21st,  a  rate  of  10  cu.  ft.  per  sec.  per  sq.  mile  has 
been  used.  This  rate  is  based  on  the  rates  in  the  basins  to  the  north 
and  south,  and  is  believed  to  be  quite  conservative. 

Molcehimne  River. — Mokelumne  Eiver  has  a  very  narrow  and  very 
long  drainage  basin  which  extends  eastward  to  the  summit  of  the 
Sierras.  From  the  low  foot-hills  to  the  junction  of  the  three  branches, 
about  30  miles  above,  its  basin  is  a  broad  canyon  with  a  minimum 
width  of  1.1  miles.  A  large  percentage  of  the  upper  basin  ranges  from 
7  000  to  9  000  ft.  in  elevation,  and  several  peaks  are  more  than  10  000 
ft.  high. 


TABLE  15. — Floav  of  Mokelumne  River,  near  Clements. 


Date,  If  07. 

Gauge  height. 

Discharge,  in  cubic  ten 
per  second. 

March  16th 

r.l5 
13.60 
17.00 
21  00 
17.9 
15.9 
13.0 
13.3 
13.0 
11.6 
11.8 

2  000 

ITth 

8  600 

18th 

12  200 

"       19th 

17  000 

"       20th 

13  000 

31st            

11200 

asd 

8  000 

2SA        

SHOO 

24th 

8  000 

25th    

6  500 

26th 

6  400 

Period. 

Mean  daily  discharge,  in  cubic 
feet  per  second. 

Total  run-off.  in  acre-feet. 

February  

2  920 
5  320 
13  350 

162  000 

327  000 

March  18th-21st 

106  000 

This   river   drains    an   area   of   about   660   sq.    miles   between  the 

American   and   Stanislaus  Rivers,   nnrl   empties  into  the  San  Joaquin 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS 


133 


cu.  ft.  per  sec.  per  sq.  mi*le.    It  is  seen  that  the  maximum  run-off  at 
tliis  station  in  1904  was  only  57%  of  that  in  1907. 

TABLE  18. — Flow  of  Stanislaus  Eiver,  at  Knights  Ferry. 


Date,  1907. 

Gauge  height. 

Discharge,  in  cubic  feet 
per  second. 

March  16th 

8.85 
14.50 
17.35 
35.30 
19.10 
15.60 
14.55 
14.15 
13.80 
14.20 
13.75 

3  780 

17th 

18th 

15  580 
24  100 

19th 

54  300 

"        30th 

31400 

"        21st 

19  300 

22cl 

15  740 

23d 

14  470 

24th 

13  430 

25th 

14  630 

26th 

10  420 

Period. 

Mean  daily  discharge,  in 
cubic  feet  per  second. 

Total  run-off,  in  acre-feet. 

February  

3  440 
9  880 
32  250 

191  000 

March 

608  000 

March  l8th-31st 

256  000 

TABLE  19. — Flow  of  Tuolumne  Kiver,  at  LaGrange. 


Date,  1907. 

Gauge  height. 

Discharge,  in  cubic  feet 
per  second. 

March  16th 

17th 

6.55 
11.30 
13.50 
15.75 
13.00 
11.50 
10.50 

9.80 
10.65 
10.65 

9.30 

3  430 
20  300 

"        18th 

33  400 

19th 

30th 

'•        21st 

51  800 
30  500 
21  500 

22d 

16  700 

38d 

13  500 

341h 

17  000 

25th 

17  000 

36th 

11  500 

Period. 

""Zl^tltf^^'^nt          Total  run-Off,  in  acre-feet. 

February 

3  910                                                217  000 

March 

11  100                                                683  000 

March  18th-21st 

34  300                                                271  000 

Tuolumne  River. — Tuolumne  River,  which  drains  an  area  imme- 
diately south  of  the  Stanislaus  River,  heads  in  the  high  peaks  of  the 
Sierras  above  Yosemite  National  Park,  at  an  elevation  of  about  13  000 
ft.,  and  empties  into  the  San  Joaquin  River  about  10  miles  west  of 


134  THE  FLOOD  OF  MAKCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

Modesto.  The  area  above  the  gauging  station  at  LaGrange  is  1  500  sq. 
miles.  Table  19  contains  data  on  the  flow  at  this  station  during  the 
1907  flood. 

The  greatest  daily  rate  of  run-off  at  this  station  during  this  flood 
was  34.5  cu.  ft.  per  sec.  per  sq.  mile,  and  the  greatest  4-day  mean  rate 
was  22.9  cu.  ft.  per  sec.  per  sq.  mile. 

Merced  River. — Merced  River  drains  the  area  between  Tuolumne 
River  and  the  Upper  San  Joaquin,  and  empties  into  the  latter  about 
26  miles  northwest  of  Merced.  It  heads  at  the  summit  of  Mt.  Lyell, 
at  an  elevation  of  13  090  ft.,  and  drains  the  southern  and  western 
slopes  of  this  mountain,  while  the  Tuolumne  drains  the  northern  slope. 
In  this  basin  is  the  famous  Yosemite  Valley,  with  its  great  waterfalls 
and  barren  domes.  The  gauging  station  on  this  stream  is  at  Merced 
Falls,  above  which  the  drainage  area  is  1  090  sq.  miles.  Table  20  con- 
tains data  on  the  flow  of  this  stream  at  the  station  during  the  flood 
of  1907. 

TABLE  20. — FLOW"  of  Merced  Kiver,  at  Merced  Falls. 


Date,  1907. 

Gauge  height. 

Discharge,  in  cubic  feet 
per  second. 

March  16th 

"     17th 

10.85 

15.2 

14.8 

18.0 

16.05 

14.8 

13.95 

13.. 55 

16.55 

15.60 

13.65 

2  200 
14  400 

'•     ISth 

'•     19th 

13  000 
23  000 

"     20th 

"     21st 

"    22(1 

17  400 
13  000 
10  200 

"     23d 

8  800 

"     24th 

"     25th                              

19  200 
15  800 

'•     26th 

9  200 

Period. 

February 

VI  arch 

March 'iStli-aist '.'.'.' 


Mean  daily  discharge,  in 
cubic  feet  per  second. 


1920 
7  170 
16  600 


Total  run-off,  in  acre-feet. 


107  000 
441000 
132  000 


The  greatest  daily  rate  of  flow  during  this  flood  was  21.1  cu.  ft. 
per  sec.  per  sq.  mile,  and  the  greatest  4-day  mean  was  15.7  cu.  ft.  per 
sec.  per  sq.  mile.  The  small  run-off  per  square  mile  arises  from  the 
fact  that  much  of  the  basin  has  a  high  altitude,  and  that  the  precipi- 
tation was  not  as  heavy  as  in  the  basins  to  the  north. 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  135 

Flow  through  Sacramento  and  San  Joaquin  Valleys. 

The  rate  of  inflow  into  the  Sacramento  and  San  Joaquin  Valleys 
from  the  metered  mountain  and  foot-hill  areas  during  this  flood  can 
he  seen  from  the  preceding  pages.  For  ready  reference,  however,  these 
rates  of  inflow  at  gauging  stations  for  the  4-day  period,  March  18th- 
21st,  are  given  in  Table  21. 

TABLE  21. — Eun-Off  from  Sacramento  and  San  Joaquin  Basins, 
IN  Cubic  Feet  per  Second,  for  March  18th-21st,  1907. 


Place. 

Drainage, 

in  square 

miles. 

Date,  Ma 

RCH,  1907 

Mean  for 
March 

18th-21st. 

18th. 

19th. 

20tb. 

21st. 

Sacramento 

Stony 

Red  BluflE 

Fruto        ... 

9  300 

601 

3  640 

1220 

263 

1  910 

1230 

805 

3  907 

4  250 

118  000 
35  000 

107  900 
85  000 
15  500 
63  200 
13  500 
19  800 

164  000 
20  000 
129  6110 
100  000 
28  000 
93  000 
19  000 
24  700 

192  000 
13  450 
84  900 
60  000 
17  400 
77  000 
13  500 
10  000 

132  000 

6  800 
66  740 
27  000 

8  400 
65  OCO 

7  820 
5  460 

151500 
16  310 

Feather 

Yuba 

Oroville 

Smarts  ville 

97  290 
68  000 

Bear 

Van  Trent 

17  300 

American 

Fair  Oaks 

74  600 

Yolo 

13  200 

Puta  .....   

Winters 

15  000 

Unmeterert  mo 
Sacramento  Vf 

antain  and  foot-hills, 
illey 

76  000* 

25  soot 

Total,  Sacramento  Basin 

27126 

554  700 

Cosumnes [Michigan  Bar. 

Mokelumne  . . .  Clements. 

Calaveras 

Stanislaus 

Tuolumne 

Merced .  . 

San  Joaquin. .. 


Jenny  Lind 

Knights  Fe^ry 

LaGrange 

Merced  Falls 

Pollasky 

Unmetered  mountain  anJ  foot-hills, 
San  Joaquin  Valley , 


Total,  San  Joaquin  Basin 16  372 


524 

7  600 

82  600 

9  300 

3  900 

642 

12  200 

17  000 

13  000 

11  200 

395 

3  800 

26  100 

3  800 

3  300 

935 

24  100 

54.300 

31  400 

19  300 

1  500 

33  400 

51800 

30  500 

21  500 

1  0'.10 

13  000 

23  000 

17  4"0 

13  000 

1  640 
5  656 

5  890 

16  372 

13  3,50 
13  3.50 
9  250 
32  250 
34  300 
16  600 
16  40011 
67  900+ 
23  560§ 


226  960 


*  Run-off  per  square  mile  assumed  as  50%  of  precipitation  for  period,  March  17th-20th, 
or  20  cu.  ft.  per  sec. 

+  Run-off  per  square  mile  assumed  as  50%  of  precipitation  for  perifid,  March  17th-20th, 
or  12  cu.  ft.  per  sec. 

i  Run-off  per  square  mile  assumed  as  40%  of  rainfall  for  period,  March  17th-20th.  or 
6  cu.  ft.  per  sec. 

§  Run-off  per  square  mile  assumed  as  40%  of  rainfall  for  period,  March  17th-20th,  or 
4  cu.  ft.  per  sec. 

I'  Run-off  per  square  mile  assumed  as  10  cu.  ft.  per  sec. 

From  Table  21  it  is  seen  that  the  mean  rate  of  run-off  from  the 
metered  area  of  the  Sacramento  Basin  (83%  of  all  mountains  and 
foot-hills)  for  the  4-day  period,  March  18th-21st,  was  about  453  000 
cu.  ft.  per  sec.  The  estimated  run-off  for  this  period  was  76  000  cu. 
ft.  per  sec.  from  the  unmetered  mountains  and  foot-hills,  and  25  500 


136  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

cu.  ft.  per  sec.  from  Sacramento  Valley,  making  a  mean  rate  of  run- 
oif  from  the  Sacramento  Basin  of  about  555  000  cu.  ft.  per  sec.  for  4 
consecutive  days. 

It  is  not  possible  to  trace  the  movement  of  this  v^ater  through  the 
valley,  on  account  of  overflow  into  flood  basins  and  breaks  in  the  levee 
system.  The  levees  failed  at  many  places  on  both  sides  of  the  Sacra- 
mento Eiver,  and  also  on  some  of  its  tributaries,  and  it  is  impossible 
to  compute  the  flow  through  any  of  these  breaks.  Such  an  estimate, 
if  correctly  made,  would  have  practically  no  value,  as  it  would  give 
little  idea  of  the  distribution  of  flow  through  the  valley  during  any 
other  flood  when  failure  of  levees  occurred  at  other  places. 

While  an  estimate  of  the  volume  passing  specified  places  in  the 
valley  at  a  given  time  cannot  be  made,  the  points  where  large  volumes 
left  the  channel  and  returned  to  it  again  or  crossed  it  can  be  indi- 
cated, as  well  as  the  time  of  failure  of  important  levees.  On  the 
evening  of  March  20th  the  water  was  overtopping  the  levees  for  almost 
the  entire  distance  between  Princeton  and  Jacinto,  and  also  above  and 
below  Colusa.  On  March  21st,  eleven  breaks  in  the  levees  occurred 
between  Colusa  and  Grimes,  and  during  that  night  several  breaks 
occurred  in  the  levees  on  the  east  side  of  Sacramento  River  between 
Clarksburg  and  Courtland,  allowing  water  from  the  Sacramento  to 
pass  into  Mokelumne  River  and  thence  into  the  San  Joaquin.  On 
March  22d  several  other  breaks  occurred  in  Colusa  County,  and  also 
in  the  Island  District,  where  large  areas  of  reclaimed  land  were  sub- 
merged. On  March  23d  the  levees  of  Ryer,  Tyler,  Brannan,  Andrus, 
and  Bouldin  Islands  and  the  Lisbon  District  failed,  flooding  65  000 
acres  of  land.  Besides  the  failures  already  mentioned,  there  were 
numerous  others  of  more  or  less  seriousness  in  different  places  in  the 
Sacramento  Valley. 

At  Knights  Landing,  on  March  21st,  the  Sacramento  was  1  ft. 
higher  than  recorded  at  any  previous  time.  Below  this  point,  a  large 
part  of  the  water  from  the  Feather  River  was  flowing  across  the  Sacra- 
mento Channel  into  Yolo  Basin.  Through  the  Kripp  crevasse  of 
February  8th,  opposite  the  City  of  Sacramento,  a  large  part  of  the 
waters  of  the  Sacramento  and  American  Rivers  also  passed  into  Yolo 
Basin,  and  the  water  level  of  this  basin  was  several  feet  higher  than 
ever  known  before.  On  February  24th  the  Sacramento  at  Rio  Vista 
reached  its  greatest  height  during  the  flood,  being  3  ft.  higher  than 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  137 

indicated  by  previous  records.  The  failure  of  the  levees  of  Brannan, 
Twitchell,  and  Andrus  Islands,  near  the  mouth  of  Cache  Slough,  per- 
mitted a  part  of  the  water  of  Yolo  Basin  to  flow  across  the  Sacramento 
Channel  into  the  San  Joaquin  River,  submerging  large  areas  in  the 
San  Joaquin  Delta.  In  all,  it  is  estimated  that  about  300  000  acres 
of  reclaimed  land  were  submerged  during  this  flood.  Below  the  City 
of  Sacramento,  the  only  reclaimed  districts  having  levees  that  with- 
stood the  high  waters  are :  Reclamation  District  No.  744 ;  Merritt 
Island,  Grand  Island,  and  Randall  Island  Reclamation  Districts; 
Geo.  W.  Locke,  private  reclamation;  Reclamation  District  No.  545; 
Sutter  and  Sherman  Islands;  and  the  northern  portion  of  Union 
Island. 

Referring  again  to  Table  21,  it  is  seen  that,  in  all  the  streams  of 
the  San  Joaquin  Basin,  the  greatest  rate  of  flow  occurred  on  March 
19th.  On  this  date  the  mean  rate  of  run-ofl  from  the  metered  area 
(41%  of  all  mountains  and  foot-hills)  was  about  205  000  cu.  ft.  per 
sec.  The  rate  from  the  unmetered  area  must  have  been  at  least 
84  000  cu.  ft.  per  sec.  from  mountains  and  foot-hills  and  24  000  cu.  ft. 
per  sec.  from  the  valley,  making  a  maximum  run-off  of  about  313  000 
cu.  ft.  per  sec.  from  the  San  Joaquin  Basin.  The  mean  rate  for  4 
days,  March  18th-21st,  was  about  227  0)0  cu.  ft.  per  sec.  It  is  im- 
possible to  indicate  the  volume  of  flow  at  different  points  in  this  valley 
owing  to  the  failure  of  levees  on  both  the  San  Joaquin  and  Sacra- 
mento Rivers,  and  the  passage  of  a  large  volume  from  the  latter  into 
the  former,  producing  back-water  and  retardation  nf  flow. 

It  is  also  seen  from  Table  21  that  the  mean  flow  from  the  mountains 
and  foot-hills  of  the  Sacramento  and  San  Joaquin  Basins  combined, 
for  the  4  days,  March  18th-21st,  was  about  732  000  cu.  ft.  per  sec.  It 
is  seen,  too,  that  the  mean  rate  of  discharge  into  Suisun  Bay  for 
these  4  days,  if  storage  in  the  valleys  had  not  been  permitted,  would 
have  been  about  782  000  cu.  ft.  per  sec,  a  volume  for  these  4  days  of 
G  200  000  acre-ft.,  or  9  690  mile-ft.,  enough  to  cover  both  basins  to 
a  depth  of  2.56  in.,  if  spread  over  them  evenly. 

Table  22  shows  the  run-off,  expressed  as  depth,  in  inches,  over  the 
drainage  basin,  together  with  the  precipitation  for  the  March  flood. 
Of  course,  there  is  the  very  regrettable  condition  of  too  few  and  poorly 
placed  precipitation  stations,  but  it  is  believed  that  the  records  here 
given   are   quite   representative   for   the    different    basins.      This   table 


138  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 


gives  some  idea  of  the  effects  of  altitude  and  of  melting  snow  in  the 
various  drainage  areas. 

TABLE  22. — KuN-OFF,  as  Depth,  in  Inches. 


Place  of  gauging. 

1 

oi 

cS    • 
01  a 

□ 

1 
o 

P 
■0.9 

Run-off  per  Square  Mile,  March,  1907. 

o 

P  a 
.S| 

Mean  for  March  18th-21st. 

stream. 

a 

il 

tj_i  o 
o  ® 
3 

u 

8.3 
35.5 
16.3 
26.7 
13.2 
55.7 
75.3 
39.1 
27.1 

9.3 
18.6 
25.5 
20.8 
23.4 
34.5 
22.9 
15.7 
10.0 

.s  • 

p' 
a<p 

Pit 

Bieber 

(Gregory 

Red  Bluff 

Oroville 

Crescent  Mills. . 

Smartsville 

Van  Treut 

Fair  Oaks 

Fruto 

Yolo 

Winters 

Michigan  Bar  . . 

Clements 

Jenny  Lind 

Knights  H  erry. . 

La(irange 

Merced  Falls... 
Pollasky 

2  950 
(iOS 

9  300 

3  640 
740 

1  220 

263 

1910 

601 

1  230 

805 

524 

642 

395 

935 

1500 

1  090 

1640 

9  900 

14  400 

14  400 

10  000 

7  000 

9  000 

5  500 
9  600 

"Vmo 

10  000 

6  000 
11500 
13  000 
13  000 
13  000 

9.3 
50.0 
20.7 
35.6 
15.5 
82.0 

106.5 
48.7 
41.6 
15.5 
39.1 
62.2 
26.5 
66.2 
58.1 
34.5 
21.1 

(Est.) 

1.23 
5.28 
2.43 
3.97 
1.96 
8.29 
11.30 
5.81 
4.03 
1.38 
2.77 
3.80 
3.08 
3.48 
5.14 
3.40 
2.34 
1.49 

McCloud 

Sacramento 

Feather 

6.56 
10.40 
10.00 
10.33 
8.13 
8.63 
5.27 
5.00 
5.10 
7.50 
6.42 
5.06 
6.26 
6.65 
5.92 
5.00 

37 

29 

Indian  Cr 

Yuba 

20 

80 

Bear 

American 

Stony  Cr 

139 
67 

76 

Cache  Cr 

Puta  Cr 

28 
54 

Cosumnes 

Mokelumna  .    ... 

Calaveras 

Stanislaus 

Tuolumne 

Merced 

51 

48 
69 
82 
51 
40 

San  Joaquin 

30 

Rate  of  Flow  in  Sacramento  Valley. 

It  will  he  instructive  to  compute  the  probable  rate  of  flow  of  the 
Sacramento  River  during  this  flood  at  the  four  places  where  it  re- 
ceives large  volumes  of  water  from  tributaries,  namely,  just  below  the 
mouths  of  Stony  Creek,  Feather  and  American  Rivers  and  Cache 
Slough,  taking  into  account  the  time  required  for  the  water  to  pass 
from  the  gauging  stations  to  the  Sacramento  and  the  time  to  pass  be- 
tween the  above-mentioned  places.  No  great  degree  of  refinement 
will  be  attempted,  as  the  data  will  not  warrant  it. 

As  a  flood  wave  travels  down  a  channel  there  is  a  gradual  diminu- 
tion of  its  height,  due  to  the  filling  of  the  channel  and  the  flattening 
of  the  wave.  Such  diminution  would  have  been  small  for  this  flood, 
and  is  neglected  in  the  computations,  for  the  following  reasons : 

(1). — The  flood  wave  was  a  long  one,  the  water  at  some  of  the 
stations  continuing  to  rise  for  4  days; 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  139 

(2). — The  streams  had  reached  a  comparatively  high  stage  on 
March  17th,  and  consequently  their  channels  were  from 
more  than  half  to  two-thirds  full  at  the  date  when  the 
computations  begin; 

(3). — The   rates  of  flow  computed  at  gauging  stations   are  24- 
hour  means,  not  maxima  for  a  few  hours. 
It  can  be  shown  that  the  speed  of  a  flood  wave,  M,  in  a  stream 


channel,  is  given  by  the  equation. 


d  Q 
dh 


M  W,  in  which  d  Q  is  the 


increment  of  discharge  corresponding  to  the  increment  of  stage,  d  h, 
and  W  is  the  channel  width.  The  value  of  M  has  been  computed  at 
each  gauging  station  for  intervals  of  1  ft.  in  gauge  height  during  the 
flood  stages,  and  a  mean  value  obtained  for  the  distance,  in  ho.urs, 
from  the  gauging  station  to  places  along  the  Sacramento  River.  These 
results  are  given  in  Table  23 : 

TABLE  23. — Data  on  Kate  of  Progress  of  Flood  Wave,  in 
Streams,  if  Water  Were  Confined  in  Channels. 


Place  to  place. 


Gauging  Station,  Sacramento  River  to  mouth  of  Stony  Creek... 

Gauging  Station,  Stony  Creeli  to  mouth  of  Stony  Creek , 

Gauging  Station,  Featlier  River  to  mouth  of  Feather  River 

Gauging  Station.  Yuba  River  to  mouth  of  Feather  River 

Gaugmg  Station,  Bear  River  to  mouth  of  Feather  River 

Mouth  of  Stony  Creek  to  mouth  of  Feather  River 

Gauging  Station,  American  River  to  mouth  of  American  River. 

Mouth  of  Feather  River  to  mouth  of  American  River 

Mouth  of  American  River  to  mouth  of  Cache  Slough 

Gauging  Station,  Cache  Creek  to  mouth  of  Cache  Slough 

Gauging  Station,  Puta  Creek  to  mouth  of  Cache  Slough 


^ 

a>  b 

>■  a> 

^ 

cS  O. 

5 

1-2 

9 

40 

35 

6 

60 

7 

50 

8 

15 

5 

100 

7 

15 

5 

ao 

7 

46 

7 

45 

4 

45 

4 

Note:  The  rate  of  travel  for  flood  waves,  as  given  above,  is  thf  mean  of  the  computed 
rates  on  each  of  the  days,  March  17th-31st,  reduced,  in  most  instances,  by  a  considerable 
percentage. 

A  study  of  the  daily  rate  of  discharge  of  the  streams  in  the  Sacra- 
mento Basin,  for  March  18th-21st,  Table  21,  shows  that  the  discharge 
at  places  along  the  Sacramento  River  was  undou'  tr'dly  at  a  maximum 
when  the  crest  of  the  wave  from  the  Feather  River  reached  them. 
This  wave  crested  at  Oroville  about  1  A.  M.,  March  19th.  As  Oroville 
is  about  9  hours  above  the  mouth  of  Feather  River,  the  crest  would 
reach  the  Sacramento  River  at  about  10  a.  m.,  March  19th,  with  a 
discharge  of  about  258  000  en.   ft.  per  sec,  including  the  Yuba  and 


140  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

Bear  Elvers.  This  amount,  combined  with  the  flow  in  the  Sacramento 
at  that  time,  would  give  the  maximum  discharge  just  below  the  mouth 
of  the  Feather  Elver.  The  flow  in  the  Sacramento  at  this  time,  how- 
ever, was  the  flow  at  the  gauging  station  above,  about  19  hours  before, 
combined  with  the  flow  at  the  gauging  station  on  Stony  Creek, 
about  20  hours  before,  or  the  flow  of  the  two  at,  say,  2  p.  M.,  March 
18th.  This  flow  was  143  000  cu.  ft.  per  sec,  which,  added  to  the 
258  000  cu.  ft.  per  sec.  from  the  Feather  Elver,  would  give  a  discharge 
of  401000  cu.  ft.  per  sec.  in  the  Sacramento.  This  volume  would 
reach  the  mouth  of  American  Elver  3  hours  later,  and  be  augmented 
by  93  000  cu.  ft.  per  sec.  passing  the  gauging  station  3  hours  before, 
sc  that  the  maximum  discharge  in  the  Sacramento  below  the  mouth 
of  the  American  Elver  would  be  about  494  000  cu.  ft.  per  sec,  and 
would  occur  at  about  1  P.  M.,  March  19th.  This  volume  would  reach 
the  mouth  of  Cache  Slough  at  about  8  p.  m.,  March  19th,  to  be  in- 
creased by  the  flow  of  the  Cache  and  Puta  Creeks  at  the  gauging 
stations  11  hours  before,  which  amounted  to  about  44  000  cu.  ft.  per 
sec.  Below  the  mouth  of  Cache  Slough,  therefore,  the  discharge  would 
have  been  about  538  000  cu.  ft.  per  sec.  It  is  to  be  noted  that  the 
maximum  flow  in  the  Sacramento  below  the  mouth  of  Stony  Creek 
was  about  205  000  cu.  ft.  per  sec,  and  did  not  occur  until  some  time 
on  March  20th. 

The  figures  just  given  do  not  include  the  unmetered  flow  of  76  000 
cu.  ft.  per  sec  from  the  mountains  and  hills  below  the  metered  basins, 
nor  the  25  500  cu.  ft.  per  sec  from  the  valley.  It  is  evident  that,  un- 
less stored  in  the  flood  basins,  it  must  have  appeared  in  the  Sacramento 
below  Cache  Slough.  It  is  impossible  to  compute  the  increase  in  dis- 
charge at  the  diflerent  places  on  the  Sacramento  Elver  due  to  these 
two  rates  of  inflow,  because  it  is  not  known  at  what  points  all  these 
waters  were  delivered;  but  it  is  quite  clear  that  there  must  have  been 
a  very  decided  increase  above  the  mouth  of  Stony  Creek  from  each 
side  of  the  river.  On  the  east  side  there  are  1  600  sq.  miles  of  moun- 
tains and  foot-hills  lying  between  the  Feather  and  Upper  Sacramento 
Basins,  which  are  drained  by  numerous  creeks,  the  most  important 
of  which  are  Mill  and  Deer  Creeks,  the  headwaters  of  which  come  from 
Lassen  Peak,  more  than  10  000  ft.  in  altitude.  Several  of  the  stations 
reporting  the  greatest  precipitation  in  March,  1907,  are  in  this  area  or 
very    near    it.      Taking    into    consideration    its   position    between   two 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS 


141 


basins  in  which  the  rate  of  i^in-ofE  is  known,  together  with  its  heavy 
precipitation  and  generally  lower  altitude,  it  is  believed  that  the  mean 
rate  of  run-off  may  be  safely  placed  at  25  cu.  ft.  per  sec.  per  sq.  mile 
for  the  period,  March  18th-21st.  This  means  40  000  cu.  ft.  per  sec. 
from  this  side.  On  the  west,  above  the  Stony  Creek  Basin,  are  1  080 
sq.  miles  of  mountains  and  foot-hills,  for  which  it  is  safe  to  put  the 
run-off  at  15  cu.  ft.  per  sec.  per  sq.  mile,  or  a  mean  of  16  000  cu.  ft. 
per  sec.  for  jVIarch  18th-21st.  This  would  mean  an  increase  in  the 
discharge  below  Stony  Creek  of  about  56  000  cu.  ft.  per  sec. 

A  considerable  area  of  mountains  and  foot-hills  between  the  Feather 
and  Bear  Basins  must  have  contributed  a  large  volume  to  the  Sacra- 
mento through  the  Feather  River,  so  that,  all  told,  the  maximum  dis- 
charge below  the  mouth  of  the  Feather  River  was  probably  at  least 
65  000  cu.  ft.  per  sec.  greater  than  that  computed  above.  As  the  rates 
of  run-off  for  the  unmetered  area  of  mountains  and  valley  are  4-day 
means,  the  maximum  discharge  below  Cache  Slough  must  have  been 
about  640  000  cu.  ft.  per  sec.  This  maximum,  however,  is  only  15% 
greater  than  the  4-day  mean  flow  of  555  000  cu.  ft.  per  sec.  for  March 
18th-21st. 

It  will  be  noticed  that  the  maximum  discharge  just  below  the 
mouth  of  Cache  Slough  would  probably  occur  at  8  p.  M.,  March  19th, 
if  the  water  were  confined  in  channels.  But  the  maximum  stage  at 
Rio  Vista,  a  few  miles  below  the  mouth  of  this  slough,  actually  oc- 
curred at  11  p.  M.,  March  23d.  Overflow  and  storage  in  the  flood 
basins,  therefore,  delayed  the  arrival  of  the  flood  crest  at  Cache  Slough 
about  4  days. 

Table  24  is  a  comparison  of  maximum  rates  of  flow  of  the  Sacra- 
mento River  during  this  flood  with  those  assumed  by  the  1904  Engi- 
neering Commission,  provided  that  the  total  run-off  is  confined  be- 
tween the  levees  and  not  allowed  to  collect  in  the  flood  basins. 

TABLE  24. 


Place. 

Maximum  rate  assumed  by 

1904  Engineering  Commission. 

Cubic  feet  per  second. 

Maximum  rate  computed 
from  March,  1907,  flood. 
Cubic  feet  per  second. 

Below  mouth,  Stony  Creek 

"           "        Feather  River . . 
"           "        American  River. 
"          "       Cache  Slough... 

180  000 
190  000 
230  000 
250  000 

261  000 
466  000 
559  000 
640  000 

142  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

These  computed  rates  are  from  45  to  156%  larger  than  the  as- 
sumed rates. 

Profile  of  Flood  Wave  in  Sacramento  Kiver. 

Fig.  2  is  a  profile  of  the  flood  wave  iu  Sacramento  River  during 
March,  1907.  This  profile  merely  shows  the  greatest  elevation  of  the 
flood  plane  above  mean  sea  level  at  different  points  along  the  course 
of  the  river.  In  other  words,  the  maximum  height  attained  by  the 
flood  at  various  points  is  platted  with  reference  to  the  distance  from 
the  mouth  of  the  river  and  the  elevation  above  mean  sea  level.  An  in- 
spection of  this  profile  shows  that  the  mean  gradient  of  the  flood  plane, 
In  feet  per  mile,  between  observed  points,  decreases  quite  rapidly  from 
Red  Bluff  toward  the  mouth  of  the  river,  actually  changing  sign  below 
Walnut  Grove.  This  gradient  varies  from  — 2.41  between  Red  Bluff 
and  Munroeville,  near  the  mouth  of  Stony  Creek,  to  +  0.01  below 
Walnut  Grove.  Such  a  reversal  of  slope  would  seem  to  indicate  a 
constricted  condition  of  the  channel  near  the  mouth  of  the  river. 

A  profile  of  the  flood  wave  of  1905,  made  under  the  direction  of  the 
Commissioner  of  Public  Works  of  California,  is  also  shown  on  Fig. 
2  for  the  purpose  of  comparison.  This  profile  may  be  considered  as 
typical  of  the  usual  flood  wave  in  the  spring  of  each  year. 

Losses  Due  to  the  Flood  of  March,  1907. 

The  losses  resulting  from  this  flood  consisted  mainly  in  the  de- 
struction of  the  crops  then  growing  on  about  300  000  acres  of  land 
completely  inundated,  together  with  the  damage  done  to  a  portion  of 
the  prospective  yield  for  the  season  of  1907.  In  addition  to  this,  many 
miles  of  costly  levees  had  to  be  rebuilt  and  many  miles  more  ex- 
tensively repaired  on  account  of  overtopping  and  wind  action.  The 
railroads  suffered  heavily,  in  bridges  and  culverts  washed  out,  in  injury 
to  miles  of  roadbed,  and  in  loss  of  traffic.  The  line  from  Marysville 
to  Knights  Landing  was  closed  from  March  19th  to  May  13th.  Among 
the  larger  bridges  swept  away  or  badly  damaged  were  the  highway  and 
the  Northern  Electric  Railway  bridges  across  the  Feather  River  at 
Oroville,  the  highway  bridge  across  the  American  River  at  Fair  Oaks, 
the  highway  bridge  on  the  Mokelumne  River  near  Clements,  and  the 
bridge  on  the  Cosumnes  River  at  Bridge  House.  Three  costly  dredges 
for  mining  gold-bearing  gravel  in  the  Feather  River  near  Oroville  were 


Papers.]    XHE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS 


143 


'^  5 
9  < 


i 

Ked 
/ 

Bluff 

h 

/ 

/ 

/^ 

^^ 

. 

A 

^,v 

/ 

o 

/    / 

Mu 

nroev 

ille 

o 

/ 

/ 

/// 

,J 

1:^ 

o 

/ 

b 
c 

B 

» 
rv 

/  / 

if 

1 

Colu 

sa 

1 1 

o 

/ 
/ 

1 

b 

/ 
;  / 

£. 

Ik 

1   !t 

1 

0 
■n 

-D 

o 

1 
1 

Knig 

Its-Lj 

ndinj 

?  ^6 

^     0  -n 

>  > 

'IS 

1 
1 

E    5  r-" 

/ 
/ 

t 

^0    p3H  -n 

1 

1 

/ 

Ig 

^^     OS     1      T1 

/ 

/ 
/ 
/ 

Sacrame 

i         1 

uto 

c 

tt- 

'  /«    Freeport 

^ 

r-     r 

a 

41  Richland 
iS  Courtland 

I 

Tl 

< 

m 

+ 

b 

Walnut  Grove 

B 

Isleton 

Mouth  Cache  SI 

Kio  Vista 

mgh 

(JoUinsville 

Elevat 

ions  in 

feet  a 

jove  m 

9an  se 

I  level 

< 

z 

5 

S       1 

S           c 

:i           c 

\ 

; 

1 

1          ? 

I         Z 

^         1 

144  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 

destroyed.  Many  towns  and  villages  were  partially  inundated,  subject- 
ing the  inhabitants  to  serious  inconvenience  at  the  time  and  to  heavy 
expense  in  repairs  later.  The  greater  part  of  Stockton  was  flooded  for 
nearly  a  week,  because  of  the  failure  of  the  levees  along  Mormon 
Slough  and  Jackson  Creek.  About  half  of  Oroville  was  flooded  for 
three  days,  and  one  hundred  and  twenty-five  families  were  driven  from 
their  homes.  The  restraining  dam  on  the  Yuba  River,  14  miles  above 
the  mouth,  known  as  Barrier  No.  1,*  was  destroyed  on  the  night  of 
March  18th.  This  dam  was  built  to  hold  back  the  mining  debris  in 
the  channel  above.  With  its  destruction,  practically  all  the  debris 
restrained  by  it  (probably  amounting  to  more  than  1000  000  cu.  yd.), 
■was  transferred  to  the  channel  below.  It  is  estimated  that  the  total 
damage  resulting  from  this  flood  exceeded  $5  000  000. 

Effect  of  Mining  Debris  on  Floods. 
From  1849  to  1880  enormous  quantities  of  debris — sand,  gravel, 
and  cobbles,  the  tailings  from  hydraulic  mining — were  deposited  in 
the  upper  course  of  several  of  the  streams  on  the  eastern  slope  of  the 
Sacramento  Basin.  The  volume  of  this  debris  in  the  Yuba  River 
alone  has  been  variously  estimated  at  from  Yl  000  000  to  700  000  000 
cu.  yd.  At  the  mouth  of  the  river,  near  Marysville,  it  has  a  depth  of 
7i  ft.;  at  Dugnens  Point,  11  miles  above  the  mouth,  it  has  a  depth 
of  26  ft.,  and  at  The  Narrows,  18  miles  above  the  mouth,  it  has  a 
depth  of  84  ft.  The  gradual  elevation  of  the  flood  plane  at  Marysville, 
due  to  the  accumulation  of  debris  in  the  channel  at  this  place,  is  shown 
by  the  maximum  gauge  readings  at  Marysville  (Table  26).  The  zero 
of  the  gauge  is  the  elevation  of  low  water  in  1872. 

TABLE  26. — Maximum  Gauge  Readings  at  Marysville. 


Date. 

Gauge  height,  t 

January 

11,  lWi3 

11  ft.     6  in. 

March 

6,  1869 

15  "     11    •• 

January 

19.  1875 

15  '•       2    '■ 

April 

23,  1880 

I')  '■       3    " 

February 

34,  1881 

18  "       2    " 

December 

23, 1884 

17  "        1     '• 

January 

18,  1898 

18  '■        5    " 

March 

25,  1899 

18  "        5    •' 

February 

31,  1901 

19  '•        0    '■ 

February 

25,  1904 

20  •'       0    '• 

January 

19,  1906 

21  "        S    •' 

February 

3,  1907 

21   "        3    " 

March 

19,  1907 

33  •'        4    " 

*The  failure  of  this  structure  is  described  in  Engineering  News,  Aug.  8th,  1907, 
t  Data  furnished  by  W.  T.  Ellis,  Levee  Commissioner. 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS  145 

The  low-water  reading  on  this  gauge,  in  the  summer  of  1906,  was 
9.0  ft.  The  flood  of  1907,  however,  changed  the  low-water  channel 
from  the  right  to  the  left  side  of  the  river,  so  that  in  Augiist  the 
elevation  of  the  water  surface  could  not  be  read  at  all,  the  debris 
around  the  gauge  being  3  ft.  higher  than  the  water. 

The  failure  of  Barrier  Dam  No.  1,  on  the  Yuba  River  14  miles 
above  the  mouth,  liberated  about  1 300  000  cu.  yd.  of  debris  which 
was  deposited  in  the  bed  of  the  stream  at  varying  distances  below  the 
dam,  depending  upon  the  size  of  the  material.  The  deposition  of  this 
enormous  volume  of  material  in  the  stream  bed,  and  the  gradual  ele- 
vation of  the  flood  plane  due  to  it,  require  frequent  raising  and  widen- 
ing of  the  levees  along  the  river.  Such  a  condition  is  fraught  with 
growing  peril  to  the  valley  land  and  to  all  interests  adjoining  the  river. 

Effect  of  Storage  Reservoirs  on  Floods. 

Any  rational  system  of  reclamation  for  the  overflow  lands  in  the 
Sacramento  and  San  Joaquin  Valleys  must  make  provision  for  passing 
the  peak  of  the  floods  rapidly  to  Suisun  Bay.  The  volume  of  flood 
water  to  be  passed  in  Sacramento  Valley,  as  determined  by  actual 
gaugings  of  the  flood  of  March,  1907,  largely  exceeds  all  estimates 
previously  used  as  a  basis  for  the  computation  of  proper  channel 
capacity  to  carry  safely  the  flood  waters  of  the  Sacramento  River. 
Indeed,  it  may  be  that  the  task  of  rectification  and  enlargement  of 
channel  necessary  to  pass  such  floods  as  that  of  March,  1907,  is  so 
great  as  to  make  it  economically  impossible.  In  such  event,  some 
auxiliary  system  of  flood  control  would  have  to  be  devised.  Probably 
no  more  effective  and  easily  executed  auxiliary  system  could  be  found 
than  that  of  large,  regulating  storage  reservoirs  in  the  mountains. 
Such  reservoirs  could  be  utilized  to  store  water  during  floods,  thereby 
reducing  the  peak  of  the  flood  in  the  valley  sufficiently  to  allow  the 
main  channel  to  carry  it  safely  to  Suisun  Bay. 

The  United  States  Reclamation  Service  has  located  the  principal 
reservoir  sites  in  the  Sacramento  Basin,  and  has  made  surveys  to 
determine  the  capacity  and  probable  cost  of  most  of  them.  Of  the 
reservoirs  surveyed  to  date,  four  are  in  Stony  Creek  Basin,  with  a 
total  capacity  of  124  100  acre-ft. ;  two  are  in  Cache  Creek  Basin,  with 
a  total  capacity  of  176  500  acre-ft. ;  two  are  in  Puta  Creek  Basin,  with 
a  total  capacity  of  318  000  acre-ft. ;  seven  are  in  Feather  River  Basin, 


146 


THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [Papers. 


with  a  total  capacity  of  775  600  acre-ft. ;  four  are  in  Pit  River  Basin, 
one  of  which  has  a  capacity  of  3  196  000  acre-ft. ;  and  one  is  on  the 
Upper  Sacramento  River  at  Iron  Canyon,  with  a  capacity  of  226  900 
acre-ft.  In  the  San  Joaquin  Basin  no  reservoir  sites  have  been  located 
and  surveyed  yet,  although  it  is  probable  that  the  area  contains  some 
good  ones. 

Table  27. — Rkservoir  Data. 


Capacity, 

in 
acre-feet. 

Drain- 
age, in 
square 
miles. 

Volume  Available  for  Storage, 
in  Acre-Feet.* 

Name  of  Reservoir. 

5 

5 

Stony  Creek  Basin. 

East  Park 

Stony  Ford 

26  000 
40  000 
14  400 
43  700 

114 

no 

58 
323 

9  390 
9  060 
4  770 
26  600 

10  600 
2  680 

13  500 

7  520 

7  250 

3  820 

21300 

5  060 

4  880 

2  570 

14  300 

2  560 
2  470 
1300 
7  250 

24  530 
23  660 
12  460 

Mill  Site      

69  450 

Cache  Creek  Basin. 

100  000 
76  500 

486 
123 

631 

14  900 
3  770 

24  600 

9  790 

2  480 

12  500 

3  460 

14  800 

998 

2  040 
31700 
23  500 

1  350 

8  000 
47  000 

55  800 

6  120 
1550 

7  820 

41  410 

Little  Indian 

Below   reservoirs   and    above 

10  460 
52  820 

PuTA  Creek  Basin. 
Guenoc  

188  000 
130  000 

148 
603 

54 

7  210 
29  400 

1980 

8  990 
36  600 

2  460 

1990 
8  110 

546 

31830 
88  910 

Below    reservoirs   and   above 

5  984 

Feather  River  Basin. 

61800 

12  600 

■     500  000 

46  270 
86  100 
68  800 

44 
682 
506 

29 

172 

1010 

1  300 

2.580 
40  000 
■29  700 

1  700 
10100 
59  300 

70  400 
49  (iOO 

3  100 
48100 
35  600 

2  040 
12  100 
71  200 

84  500 

1  600 
24  800 
18  400 

1  050 
6  260 
36  700 

436  0 

9  320 

144  600 

Big  Meadow  

107  300 

Buck's    Valley    and    Spanish 
Ranch 

6140 
36  460 

214  200 

Below  reservoirs   and    above 

254  300 

3  196  000 

Pit  River  Basin. 
Big  Valley 

3  950 

54  500 

49  600 

41200 
220  800 

194  900 

Sacramento  River. 

226  900 

6  350 

184  400 

270  500 

331400 

1007100 

*  The  daily  run-ofif  per  square  mile  is  assumed  to  be  constant  over  the  basin  abov  e  the 
gauging  station. 

In  Table  27  are  shown  the  reservoir  sites  in  the  Sacramento  Basin 
which  could  be  used  for  flood  control,  together  with  the  drainage  area 


Papers.]    THE  FLOOD  OF  MARCH,  1907,  IN  CALIFOKNIA  RIVERS  147 

tributary  to  each  and  its  capacity  in  acre-feet.  Assuming  the  run-oS 
per  square  mile  to  be  constant  in  any  particular  basin,  the  quantity  of 
water  available  for  storage  at  each  reservoir  is  given  for  each  of  the 
days,  March  18th-21st,  and  also  the  total  for  the  4  days.  It  will  be 
noted  that  some  of  these  reservoirs  would  be  only  partially  filled  by 
the  flood  flow  of  March  18th-21st,  while  others  would  store  but  a  small 
percentage  of  the  run-off  for  this  period. 

A  study  of  Table  27  will  show  that  the  four  reservoirs  in  Stony 
Creek  Basin  would  have  stored  the  run-off  from  481  sq.  miles,  or  80% 
of  the  area  above  the  gauging  station,  and  would  have  reduced  the 
maximum  daily  flow  from  25  000  to  5  000  cu.  ft.  per  sec.  The  two 
reservoirs  in  Cache  Creek  Basin  would  have  stored  the  flow  from  609 
sq.  miles,  or  50%  of  the  area  above  the  gauging  station,  and  would 
have  reduced  the  maximum  daily  flow  from  19  000  to  9  500  cu.  ft.  per 
sec.  The  two  reservoirs  in  Pvita  Creek  Basin  would  have  stored  the 
flow  from  751  sq.  miles,  or  93%  of  the  area  above  the  gauging  station, 
and  would  have  rediiced  the  maximum  daily  flow  from  24  700  to  1  700 
cu.  ft.  per  sec. 

The  seven  reservoirs  in  Feather  River  Basin  would  have  stored  the 
flow  from  about  1 134  sq.  miles,  or  31%  of  the  area  above  the  gauging 
station  at  Oroville,  leaving  2  506  sq.  miles  uncontrolled.  Of  this  un- 
controlled area,  623  sq.  miles  are  above  Mohawk  Valley  Reservoir, 
683  sq.  miles  are  above  Indian  Valley  Reservoir,  and  1  200  sq.  miles 
are  below  the  reservoirs  and  above  the  gauging  station.  This  storage 
would  have  reduced  the  daily  flow  at  Oroville  as  follows : 

From  107  900  to  74  300  cu.  ft.  per  sec.  on  March  18th ;  from  129  600 
to  89  200  cu.  ft.  per  sec.  on  March  19th;  from  84  900  to  58  500  cu.  ft. 
per  sec.  on  March  20th ;  and  from  66  740  to  45  900  cu.  ft.  per  sec.  on 
March  21st.  Big  Valley  Reservoir,  on  Pit  River,  would  have  stored  the 
entire  flow  at  that  place  and  reduced  the  daily  flow  of  the  Sacramento 
River  at  Red  Bluff  about  25  000  cu.  ft.  per  sec.  The  storage  at  Iron 
Canyon,  together  with  that  on  Pit  River,  would  have  reduced  the 
greatest  daily  flow  of  the  Sacramento  River  at  Red  Bluff  from  192  000 
to  106  000  cu.  ft.  per  sec. 

The  combined  effect  of  all  these  reservoirs  in  operation  at  the  same 
time  would  have  been  to  reduce  the  maximum  flow  in  the  Sacramento 
River  by  about  86  000  cu.  ft.  per  sec.  above  the  mouth  of  Stony  Creek, 
106  000  cu.  ft.  per  sec.  above  the  mouth  of  the  Feather  River,  and 
179  000  cu.  ft.  per  sec.  below  the  mouth  of  Cache  Slough. 


148  THE  FLOOD  OF  MARCH,  1907,  IN  CALIFORNIA  RIVERS    [P-ipeiS- 

It  would  seem  that  the  ultimate  solution  of  the  flood  problem  in 
the  lower  portions  of  the  Sacramento  Valley  is  closely  interwoven 
with  the  reclamation  of  the  higher  portions  by  irrigation.  Keservoirs 
which  would  impound  flood  waters  and  reduce  the  peak  of  floods,  so 
ap  to  save  the  lowlands  from  overflow  in  the  early  spring,  would  serve 
later  as  storage  reservoirs  from  which  to  draw  for  irrigation  purposes. 
The  flood  problem  in  this  valley  is  indeed  a  very  serious  one,  and 
merits  the  most  careful  and  thoughtful  consideration. 


Vol.  XXXIV.  FEBRUARY,  1908.  No.  2. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEEES. 

INSTITUTED    1852. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


THE  REINFOKCED  CONCRETE  WORK  OF  THE 

McGRAW  BUILDING. 

Discussion.* 


By  Messrs.  T.    L.  Condron  and    F.  F.   Sinks,  E.    W.  Stern,    L.  J. 
Mensch,  and  p.  E.  Stevens. 


T.  L.  Condron  and  F.  F.  Sinks,  Members,  Am.  Soc.  C.  E.  (by  Messrs.  Con- 
letter). — No  good  reason  can  be  offered  for  not  exercising  the  same 
common  sense  in  designing  reinforced  concrete  strilctures  as  that  ex- 
pected and  demanded  in  designing  steel  or  timber  structures.  Too 
much  is  heard  regarding  "systems"  of  reinforced  concrete,  and  too 
little  regarding  the  simple  application  of  the  well-known  hiws  relating 
to  the  strength  of  materials  and  the  distribution  of  stresses  in  such 
structures.  It  is  not  many  years  ago  that  iron  bridges  were  btiilt  ac- 
cording to  one  or  another  special,  and  generally  patented,  type.  To-day 
the  "patented  bridges"  are  limited  to  draw  bridges,  which,  after  all, 
are  machines  as  well  as  structures.  It  is  doubtless  true  that,  in  a 
large  measure,  the  present  wide  use  of  reinforced  concrete  is  due  to 
the  energetic  promotion  of  various  so-called  "systems,"  together  with 
the  equally  energetic  promotion  of  concrete  construction  by  makers 
of  various  forms  of  reinforcing  materials.  While,  in  some  cases, 
capable  and  conscientious  engineers  have  done  splendid  work  in  de- 
veloping new  and  better  designs  for  reinforced  concrete  as  a  substi- 
tute for  designs  of  steel  or  masonry  structures,  in  other  cases,  less 
capable,  and  in  some  instances  ignorant,  men  have  produced  "sys- 
tems" which  would  be  ridiculous  if  they  were  not  dangerous. 

♦Continued  from  January,  1908,  Proceedings. 


150  DISCUSSION  ON  KEINFORCED  CONCEETE  BUILDING        [Papers. 

Messrs.  Con-  In  designing  reinforced  concrete,  the  writers  have  endeavored  to 

'  follow  the  same  methods  of  analysis  of  stresses  and  proportioning  of 
parts  as  they  use  in  designing  steel  structures.  They  have  studied 
carefully  all  the  experimental  and  research  work  done  by  the  leading 
technical  schools  and  universities,  and  believe  that  more  can  be  gained 
by  such  study  than  by  simply  developing  any  refined  theoretical 
analysis  of  the  strength  of  .concrete  reinforced  with  steel. 

The  writers  present  herewith  illustrations  of  what  they  believe  to 
be  rational  designs  of  reinforced  concrete  construction.  Care  has  been 
taken  to  have  these  designs  free  from  every  unnecessary  complica- 
tion, the  whole  aim  being  to  gain  great  strength  and  everlasting 
durability  with  the  most  simple  construction  possible. 

The  author's  description  of  the  McGraw  Building  is  of  especial 
interest,  as  there  are  several  features  in  its  design  which  are  similar 
to  those  used  by  the  writers;  therefore,  they  present  the  following 
description  of  one  building,  and  some  notes  regarding  two  others, 
designed  by  them. 

The  Manufacturers'  Furniture  Exchange  Building,  in  Chicago, 
the  reinforced  concrete  features  of  which  were  designed  by  the  writers, 
as  Consulting  Engineers  for  the  Architect,  Mr.  William  Earnest 
Walker,  was  designed  in  the  spring  of  1906  and  completed  near  the 
close  of  that  year.  In  the  McGraw  Building,  as  well  as  in  the  build- 
ings designed  by  the  writers,  the  columns  have  been  reinforced  with 
latticed  steel  angles.  As  far  as  the  writers  are  aware,  the  first  rein- 
forced concrete  building  in  which  columns  of  this  form  were  used 
was  the  Watson  Building,  in  Chicago,  built  in  1905,  for  which  Messrs. 
Huehl  and  Schmidt  were  the  Architects.  The  writers'  original  recom- 
mendation for  the  columns  of  this  building  was  that  the  angles  be 
latticed,  but  they  were  actually  built  with  horizontal  tie-plates,  as 
shown  by  Fig.  1,  Plate  XXIV.  At  the  time  this  photograph  was  taken 
the  view.  Fig.  2,  Plate  XXIV,  was  also  taken  on  the  first  floor,  where 
concreting  was  going  on,  the  forms  for  the  floors  above  being  siipported 
so  that  they  did  not  interfere  with  the  placing  of  concrete  on  this 
floor. 

The  general  plans  for  the  Manufacturers'  Furniture  Exchange 
Building  were  completed  in  June,  1906,  and  the  contracts  were  let 
about  July  1st.  The  building  is  near  the  business  center  of  Chicago, 
and  has  a  frontage  of  70  ft.  on  Wabash  Avenue,  running  back  170  ft. 
on  Fourteenth  Street  to  an  alley.  The  general  appearance  of  the 
building  is  shown  by  Fig.  1,  Plate  XXV.  It  is  an  eight-story  and  base- 
ment building,  designed  for  furniture  show  rooms,  warehouse  pur- 
poses, or  light  manufacturing.  The  floors  are  designed  to  carry  live 
loads  of  160  lb.  per  sq.  ft.  on  the  lower  floors,  and  100  lb.  per  sq.  ft.  on 
the  upper  floors.  Fig.  2  is  a  plan  and  Fig.  3  a  cross-section  of  the 
building,  showing  the  general  arrangement  of  the  columns  and  beams. 


PLATE  XXIV. 
PAPERS,  AM.  SOC.   C.   E. 

FEBRUARY,   1908. 

CONDRON   AND  SINKS    ON 

REINFORCED  CONCRETE  BUILDING. 


Fig.  1.— Second  and  Third  Story  (_".ii,i  mx  Keinforcin  j,  Watson  Building. 


Fig.  2.— Concreting  on  First  Floor,  Watson  Building. 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING 


151 


On  the   first  floor   a  bulkhead  is  carried  around  on  two  sides,  sup-    Messrs.  Con- 
porting  platforms  for  the  show  windows  and  permitting  half  windows 
for  lighting  the  basement.     The  second  to  eighth  floors,  inclusive,  are 
exactly  alike.     The  roof  is  of  reinforced  concrete,  and  has  two  rows 
of  saw-tooth  skylights. 


[^ 70  iX- 


E3 


u 

levator  MA/  ^ 
Shaft-Ty\ 


[]  o  ®  e& 


=©=©= 


0 


T 


PLAN  OF  2ND  TO  8TH  FLOOR 
Fig.  2. 


4  Concrete  TaU 


Fig.  3. 


In  designing  the  columns,  the  ratio  of  the  moduli  of  elasticity  of 
steel  and  concrete  was  assumed  as  15  to  1.  The  columns  were  not 
considered  as  hooped  concrete,  only  500  lb.  per  sq.  in.  being  allowed 
for  the  working  stress  on  the  concrete  and  7  500  lb.  per  sq.  in.  on  the 
steel.  Only  one  change  was  made  in  the  size  of  the  concrete  columns. 
From  the  basement  to  the  third  story  the  columns  were  24  in.  square, 
and  above  that  they  were  20  in.  square.  The  corners  of  the  columns 
were  rounded  to  a  radius  of  4  in.,  except  in  the  basement. 

Fig.  4  shows  the  typical  reinforcement  of  the  columns,  girders,  and 
slabs.     Temporary  cross-angles  were  bolted  to  the  steel  column  rein- 


153 


DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 


Messrs.  Con- 
dron  and  Sinks. 


forcement  to  support  the  column  forms,  and,  in  turn,  the  floor  forms 
above.  After  the  concrete  for  one  floor  was  finished,  the  weight  of 
the  form  work  of  the  floor  above  was  supported  by  shores  in  the  usual 
manner,  resting  directly  on  the  finished  concrete  floor.  The  temporary 
angles  were  then  removed  from  the  columns,  and  the  column  boxing 
was  closed,  preparatory  to  casting  the  concrete  in  the  column  section 
above  the  finished  floor.  This  is  all  shown  quite  clearly  in  Figs.  1 
and  2,  Plate  XXVI. 


-I'OK-^' 


%  Finish 


Fig.  1,  Plate  XXVI,  is  a  photograph  taken  at  the  beginning  of  the 
concreting  work  on  the  first  floor  (October  10th,  1906),  and  when  taken, 
the  forms  were  completed  for  the  first  floor,  the  reinforcement  of  this 
floor   was   in    place,    the    column    reinforcement,    extending    from    the 


PLATE  XXV. 
PAPERS,   AM.   SOC.  C.   E. 

FEBRUARY,  1908. 

CONDRON  AND  SINKS  ON 

REINFORCED    CONCRETE    BUILDING. 


./f       /  /i  ■    A^ 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING  153 

footings  to  the  level  of  the  second  floor,  was  also  in  place,  and  the  ^®^,^''^:  ^P^- 
basement  columns  were  cast.     The  character  of  the  column  reinforce-  ' 
ment  is  shown  very  clearly  in  this  photograph.     The  points  where  the 
lacing  of  the  columns  is  omitted  near  the  top  are  the  openings  left 
for  the  reinforcement  for  the  second-floor  beams  to  pass  through. 

In  Fig.  2,  Plate  XXVI,  the  temporary  angles  may  be  seen  near  the 
bottom  of  the  column  in  the  foreground.  This  photograph  was  taken 
on  November  6th,  and  shows  the  concreting  in  progress  on  the  second 
floor.  After  the  concrete  work  on  the  first  floor  was  finished,  no  more 
concreting  was  done  until  after  the  forms  for  the  second  and  third 
floors  were  both  completed.  The  carpenters  then  worked  on  the  third 
floor,  building  the  forms  for  the  fourth  floor.  At  this  stage  of  the 
work  the  reinforcing  material  for  the  second  floor  was  placed,  and 
the  concreting  of  this  floor  proceeded.  The  photograph.  Fig.  1,  Plate 
XXVII,  was  taken  at  the  same  time  as  Fig.  2,  Plate  XXVI,  from  which 
it  will  be  seen  that  the  exterior  walls  had  been  run  up  practically  to  the 
level  of  the  fourth  floor,  and  the  carpenters  are  seen  working  on  the 
fourth-floor  forms,  and,  as  stated  previously,  concrete  was  being  placed 
on  the  second  floor.  From  this  time  forward,  both  the  concrete  gang 
and  the  carpenters  were  kept  constantly  at  work,  the  carpenters  being 
two  stories  ahead  of  the  concrete  gang.  In  order  to  prevent  freezing, 
the  window  openings  were  closed,  and  coke  fires  were  kept  burning  in 
salamanders  in  the  story  directly  under  the  floor  on  which  concrete 
was  being  placed.  As  a  consequence,  the  concreting  went  on  at  a 
temperature  that  was  considerably  above  the  freezing  point,  even  in 
the  coldest  weather.  The  cement  finish  was  put  on  the  concrete  floors 
as  soon  as  the  first  concrete  had  taken  its  initial  set.  Owing  to  the 
prevalence  of  rainy  weather  during  this  construction,  considerable 
trouble  was  caused  by  water  dripping  from  the  forms  on  the  newly 
finished  cement  floors,  and  great  care  had  to  be  exercised  to  protect 
these  floors  from  injury. 

Fig.  2,  Plate  XXV,  is  a  photograph,  taken  on  December  15th,  when 
the  exterior  walls  were  finished.  It  shows  the  concrete  hoist  on  the  side 
of  the  building,  with  a  dumping  bucket  just  below  the  level  of  the 
fourth  floor.  The  scaffold  at  the  rear  of  the  building  carried  the 
elevator  used  for  raising  brick.  The  concrete  mixer  is  shown  on  the 
ground  at  the  base  of  the  concrete  hoist. 

As  already  stated,  concreting  on  the  first  floor  began  on  October 
10th,  and  the  concrete  roof  over  the  eighth  story  was  completed  on 
December  27th,  only  66  working  days  intervening  between  the  time 
of  starting  the  first  floor  and  the  completion  of  the  roof.  On  Decem- 
ber 31st  the  tenants  began  moving  into  the  practically  finished  build- 
ing. 

Fig.  2,  Plate  XXVII,  is  a  photograph  of  the  third  floor,  taken  on 
December  29th,  and  is  typical  of  the  upper  floors. 


154 


DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 


Messrs.  Con- 
dron  and  Sinks. 


The  details  of  the  reinforced-concrete  construction  are  shown  quite 
clearly  in  Fig.  4.  It  will  be  seen,  both  in  the  beams  and  in  the  slabs, 
that  some  of  the  reinforcing  bars  run  straight  through  on  the  bottom, 
and  some  are  turned  up  to  form  the  reinforcement  for  the  upper  face 
of  the  beams  and  slabs  in  such  manner  as  to  provide  as  much  steel  in 
the  top  of  the  beams  and  slabs  over  the  supports  as  in  the  bottom  of 
the  beams  and  slabs  between  the  supports. 

Fig.  5  is  a  bill  of  bars  for  the  slabs  of  one  floor,  and  Fig.  6  is  the 
bill  of  bars  for  one  girder,  taken  from  the  working  plans.  These  bills 
of  bars  were  prepared  carefully  and  were  shown  on  the  working  plans, 
together  with  sketches  of  each  different  beam  or  girder,  so  that  the 


.Mkd. 

Total 
No.  of  Bars 

Size 

Length 

Shape 

Location 

A-A 

17'6" 

Y^^ 

^ 

i^,toiJ,incl. 

13^ 

'.'. 

^ ^10-6^^ H 

^ 6-0^^— ^- 

B-B 

2-M 

- 

14' 0" 

Straight 

"l9-20-22-23-24 

C-C 

9 

" 

6'0" 

" 

P., 

D-D 

20 

" 

8'0" 

•(• 

R. 

3- 

^f 

1 

E-E 

235 

" 

2'6" 

T^l< 

j 

Anchors  , 

v      1 

F-F 

5 

" 

14' 0" 

+                                                            Mo: 

Pu 

s" 

V 

•t 

-—5-0^ 

\^ 

H-H 

5 

" 

9' 6" 

straight 

I'll 

X-X 

224 

'• 

23' 0" 

" 

Longitudinal 

Y-Y 

18 

" 

18'0" 

" 

" 

H-H 

3 

" 

9 '6" 

1 

" 

J-.T 

4 

" 

IV' 6" 

10  '6  ■ 

i> T-o"      '^ 

P.. 

1 < 

BILL  OF  BARS  IN  SLABS  FOR  ONE  FLOOR 

Fig.  5. 

Contractor  was  able  to  get  out  the  correct  number  of  bars,  and  bend 
them  to  the  shape  required.  By  following  the  plans,  it  was  a  simple 
matter  to  select  the  right  bars,  and  place  them  properly  in  the  beams 
and  panels.  The  specifications  required  that  no  concrete  should  be 
put  in  until  the  inspector  had  checked  and  approved  the  placing  of  the 
reinforcing  material.  The  floor  bars  were  held  the  proper  distance 
above  the  forms  by  2-in.  lengths  of  round  iron  of  the  proper  diameter, 
while  the  bars  in  the  beams  were  supported  by  two  cement  blocks  in 
each  beam  like  those  shown  in  Fig.  4.  About  1  300  of  these  cement 
blocks  were  required  for  the  entire  building,  and  each  block  was  re- 
inforced with  two  No.  8  wires,  so  that  they  could  be  handled  safely. 
These  blocks  were  found  to  work  perfectly,  the  bars  resting  in  the 
notches,  thus  being  held  in  their  proper  places  while  the  concrete  was 
poured  into  the  beams. 


PLATE   XXVI. 

PAPERS,   AM.   SOC.   C.   E. 

FEBRUARY.   1908. 

'CONDRON   AND  SINKS  ON 

REINFORCED   CONCRETE   BUILDING. 


Fig.  1.— Reinforcing  of  Columns  and  Floor,  Manufacturers'  Furniture  Exchange 

Building. 


Fig.  2.— Concreting  on  Second  Floor,  Manufacturers'  Furniture  Exchange 

Building. 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING 


155 


Fig.   7  shows  the  reinforcement  of  the  stairways,  of  which  there    Messrs.  Con- 
were  two  flights  running  from  the  basement  to  the  eighth  floor.  ronan     m  s. 

The  extreme  simplicity  of  the  reinforcement  of  this  building  is 
evident,  and,  as  corrugated  bars  were  used,  it  was  not  necessary  to  make 
bends  at  the  ends  of  the  bars  or  use  other  means  of  insuring  bond,  the 
form  of  the  bar  giving  in  all  cases  an  absolute  bond  between  the  con- 
crete and  the  steel. 

The  total  cost  of  the  reinforcing  bars  delivered  in  Chicago  was 
almost  exactly  5%  of  the  cost  of  the  building,  and,  while  the  special 
bar  used  cost  more  than  plain  or  other  forms  of  reinforcing  bars, 
the  saving  which  would  have  been  made  by  using  a  less  expensive  one 
would  have  been  insignificant  as  compared  with  the  cost  of  the  build- 
ing. 


Jilk. 


l.-.l 


Xo.  of  Bars 
in  each  Beara 


'i  -16  0 


-21  0 


eo  0 


Shape 


Straiofht 


-5-1-1- 


9-0- 


-5-14- 


_1. 


BARS  REQUIRED  FOR  BEAMS  G^ 
Fig.  6. 

ITnder  the  Chicago  Building  Ordinances,  it  is  necessary  for  rein- 
forced-concrete  floors  to  be  tested  with  a  load  at  least  double  that  for 
which  they  are  designed,  and  this  ordinance  requires  that  the  floors  thus 
tested  shall  show  no  evidence  of  failure  and  shall  not  deflect  more  than 
T^s  of  the  span,  or  i  in.  for  a  14-ft.  span. 

ITnder  this  ordinance,  these  floors  were  tested  with  a  load  of  350 
lb.  per  sq.  ft.,  covering  an  entire  panel  of  14  by  17  ft.,  under  which 
test  load  a  deflection  of  less  than  tV  in-  was  measured. 

Later,  the  writers  followed  this  method  of  design  for  the  ware- 
house of  the  Advance  Thresher  Company,  at  Kansas  City  (Mr.  J.  C. 
Llewellyn,  Architect),  the  typical  reinforcement  of  the  columns, 
girders,  beams,  and  floor  slabs  of  which  is  shown  in  Kg.  8.  These 
floors  were  designed  to  carry  a  working  load  of  250  lb.  per  sq.  ft.  in 
addition  to  the  dead  load.     In  this  case,  floor  slabs  of  8  ft.  span  were 


156 


DISCUSSION  ON  REINFOECED  CONCRETE  BUILDING        [Papers. 


Messrs.  Con- 
dron  and  Sinks. 


adopted,  and  these  were  carried  by  concrete  joists  framing  into  con- 
crete girders.  The  spacing  of  the  columns  was  very  irregular  in  the 
building,  but  the  maximum  spans  were  24  ft.  The  type  of  column 
reinforcement  was  about  the  same  as  in  the  Manufacturers'  Furni- 
ture Exchange  Building,  except  that  the  angles  for  this  building 
were  placed  with  the  corners  out,  instead  of  in. 


FlG.  7. 

Owing  to  the  fact  that  this  building  is  to  be  used  for  exceedingly 
heavy  concentrated  loads,  that  is,  the  weight  of  traction  engines  having 
12  000  lb.  concentration  on  a  wheel,  the  owners  desired  to  satisfy  them- 
selves of  the  effect  of  concentrated  loads  on  these  floors,  and  therefore, 
the  following  test  was  made : 


PLATE  XXVII. 
PAPERS,   AM.  SOC.  C.   E. 

FEBRUARY,  1908. 

CONDRON   AND  SINKS  ON 

REINFORCED  CONCRETE   BUILDING. 


Fig.  1.— Exterior  of  Manufacturers'  Furniturf  Ex'^hange  BriLniNG.  Nov.  0th,  1906. 


l<'jc;.  3.— Typical  Interior,  Second  to  Seventh  Stories,  MANi'FACTrRERs'  Fuknituke 
Exchange  Building. 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING 


157 


Two  strips  of  2  by  4-in.  wood,  each  4  ft.  long,  were  laid  along  the    Messrs.  Con- 
centers of  two  adjoining  8-ft.  slab  spans  parallel  with  the  supporting  ''™°^'^<i^'"*'s. 
joists.     On   these   two   strips    rested   a  platform   on  which   a   load   of 
40  850  lb.  was  placed,  giving  a   concentration  of  20  425  lb.  on  each 
strip,  or  a  concentrated  load,  at  the  center  of  each  of  these  two  slabs, 
of  5  106  lb.  per  lin.  ft.,  thus  producing  the  same  moment  in  the  slab 


TYPICAL  REINFORCEMENT  FOR 

THE  ADVANCE  THRESHER  COMPANY'S  WAREHOUSE 

KANSAS  CITY, MO. 

Fig.  8. 

as  a  uniformly  distributed  load  over  a  4-ft.  width  of  the  two  panels 
of  1 274  lb.  per  sq.  ft.  Under  this  test  a  deflection  of  tV  in-  was 
measured.  Of  course,  the  entire  floor  assisted  in  carrying  such  a  test 
load,  and  this  is  only  mentioned  as  illustrative  of  the  remarkable  carry- 
ing capacity  of  such  floor  slabs.  Notwithstanding  the  fact  that  such  floor 
tests  give  astonishing  results,  the  writers  believe  that  floors  should  be 
designed,  not  on  the  basis  of  such  tests,  but  in  accordance  with  con- 


158  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Messrs.  Con-  servative  practice,  and  in  all  their  work  they  have  considered  that, 
*  for  floor  slabs  and  intermediate  beams,  where  the  reinforcement  passes 
over  the  top  of  the  slabs  and  beams  at  the  ends  and  well  into  the  next 
panel,  the  moment  is  equal  to  iv  I-  -=-  12,  and  that,  for  end  beams  and 
end  panels,  where  the  reinforcement  can  only  pass  in  this  manner 
over  the  top  at  one  support,  w  P  -^  10.  The  writers  consider  that  the 
dead  load  plus  the  assumed  live  load  will  stress  the  reinforcing  ma- 
terial to  one-third  of  its  elastic  limit,  and  they  use  from  0.8  of  1% 
to  1%  of  steel  reinforcement  having  an  elastic  limit  of  50  000  lb.  per 
sq.  in.  They  have  not  calculated  the  beams  as  T-beams,  even  in  such 
floor  construction  as  illustrated,  but  have  considered  them  as  rectangu- 
lar beams  with  a  depth  equal  to  the  distance  from  the  top  of  the  floor 
to  the  bottom  of  the  beam,  and  in  them  have  used  reinforcement  as 
great  as  li%,  but  usually  it  does  not  exceed  li%,  of  the  area  of  the 
beam,  not  including  any  portion  of  the  floor  slab  except  that  which  is 
a  part  of  the  beam  section. 

In  designing  beams  and  girders,  the  writers  have  used  the  empirical 
formula,  M  =  (450  P  +  55)  hd^.  This  formula  was  adopted  as  a 
result  of  the  study  of  all  the  tests  of  reinforced  concrete  beams  which 
bad  been  made  in  the  various  engineering  laboratories  of  the  techni- 
cal schools  prior  to  June,  1905.  Up  to  that  time,  202  beam  tests  had 
been  reported,  of  which  72  were  of  beams  reinforced  with  Johnson 
bars  having  an  elastic  limit  of  about  50  000  lb.  per  sq.  in.  Of  these 
72  tests,  80%  showed  ultimate  strengths  exceeding  that  given  by  the 
formula,  while  only  8%  developed  less  than  90%  of  the  formula 
strength,  and  the  lowest  test  developed  78  per  cent.  This  formula  is 
only  used  where  P,  the  percentage  of  reinforcement,  is  more  than 
half  of  1%,  and  not  more  than  1^%,  and  where  steel  having  an  elastic 
limit  of  50  000  lb.  per  sq.  in.  is  used,  and  with  a  positive  mechanical 
bond.  For  1%  of  reinforcement,  the  ultimate  moment,  M  =  505  td^, 
and,  in  general  that  percentage  has  been  used.  The  writers  have  con- 
sidered that  three  times  the  dead-load  moment  plus  three  times  the 
live-load  moment  is  equal  to  the  ultimate  moment.  Therefore,  if  the 
dead  load  is  equal  to  one-half  of  the  live  load,  it  would  require, 
theoretically,  an  application  of  four  times  the  working  live  load  to 
reach  the  ultimate  load.  The  tests  made  indicate  that  this  practice  is 
on  the  safe  side. 
Mr.  stern.  E.  W.  Stern,  M.  Am.  Soc.  C.  E.— The  author  states  that,  in  view 
of  the  uses  to  which  the  McGraw  Building  was  to  be  devoted,  it  was 
imperatively  necessary  that  it  should  be  designed  to  afford  the  greatest 
possible  resistance  to  the  vibration  of  heavy  machinery.  Now,  is 
enough  known  about  the  action  of  reinforced  concrete  under  vibratory 
loads  to  make  certain  its  suitability  for  this  purpose? 

In  reinforced  concrete  buildings,  cracks  occur  when  there  are  prac- 
tically no  vibratory  loads;  under  the  influence  of  vibrations  continued 


Papers.]        DISCUSSION  ON  REINFOKCED  CONCRETE  BUILDING  159 

for  a  immber  of  years,  due  to  running  machinery  in  the  building,  is  Mr.  stern, 
it  certain  that  cracks  will  not  develop,  and  that  the  reinforcing  rods 
will  not  work  loose  in  the  concrete? 

Considering  the  design  of  the  columns,  the  author  assumes  that 
stress  is  transmitted  into  the  concrete  filling  through  the  rivet  heads 
and  lattice  bars  of  these  columns,  so  that  both  steel  and  concrete  act 
together  as  a  unit.  The  speaker  cannot  accept  this  assumption.  It 
seems  to  him  far-fetched  and  entirely  problematical.  If  there  are  any 
experiments  to  fortify  the  contention  of  the  author,  it  would  be  of 
benefit  to  this  discussion  to  have  these  results. 

The  author  likewise  adopts  a  working  stress  of  750  lb.  per  sq.  in. 
on  the  concrete  filling  of  the  columns,  equivalent  to  45  tons  per  sq.  ft. 
Such  a  very  high  unit  stress  is  so  much  more  than  has  been  con- 
sidered good  practice  (being  more  than  double  that  allowed  in  the 
Building  Code  of  Manhattan),  that  the  author  should  give  the  Pro- 
fession the  benefit  of  the  experiments  upon  which  he  bases  his  con- 
clusions. 

It  is  not  clear  to  the  writer,  in  examining  the  details  of  the 
columns,  how  the  splices  of  the  columns  were  arranged  at  the  various 
joints  to  take  care  of  the  reduction  in  dimensions.  For  instance,  the 
columns  in  the  ninth  story  are  17  by  17  in.,  back  to  back  of  angles, 
whereas  the  next  section  of  columns,  supported  on  these,  is  decreased 
suddenly  to  10  by  10  in.,  back  to  back  of  angles.  It  would  be  of  interest 
to  know  how  this  change  in  size  was  taken  care  of  in  the  details. 

The  speaker  believes  the  type  of  column  used  to  be  neither  as 
economical  nor  as  efficient  as  a  box  steel  column  made  of  plates  and 
channels.  A  column  of  this  type,  to  take  the  same  load,  would  be 
made  of  15-in.  channels  with  17-in.  cover-plates,  and  would  build  up 
about  21  in.  square,  if  surrounded  with  2  in.  of  fire-proof  covering. 
The  columns  in  the  McGraw  Building  are  29  in.  square  in  the  base- 
ment and  first  floor,  or  about  40%  larger  in  outside  dimensions,  and 
occupy  nearly  twice  as  much  space;  in  fact,  in  the  lower  stories,  these 
columns  are  actually  about  as  large  in  outside  dimensions  as  the  steel 
columns  in  the  thirty-two  story  City  Investing  Building,  at  Broadway 
and  Cortlandt  Street,  designed  by  the  firm  of  which  the  speaker  is  a 
member. 

There  is  actually  more  steel  in  the  type  of  column  adopted  than 
there  would  be  in  the  channel  and  plate  box  column  above  mentioned, 
assuming  that  the  entire  load  were  carried  by  it,  without  any  regard 
to  concrete  filling;  and,  if  it  were  intended  to  fill  these  columns  solid 
with  concrete  and  surround  them  with  a  fire-proof  covering  of  that 
material,  there  would  be  less  concrete  used,  so  that  the  column  adopted 
was  extravagant  both  in  material  and  in  space  occupied. 

The  author  claims  that  the  use  of  the  type  of  steel  column  adopted 
was  a  great  convenience  in  erection,  as  it  enabled  the  steelwork  to  be 


160  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Stern,  erected  ahead  of  the  concrete  work,  and  afforded  convenient  support- 
ing members  for  the  adjoining  forms  or  for  other  erection  work. 
This  argument  would  be  equally  applicable  to  the  more  economical 
type  of  column  suggested  by  the  speaker. 

The  columns  are  spaced  too  closely  together  for  a  building  adapted 
to  loft  purposes.  The  interior  columns,  parallel  to  39th  Street,  are 
15  ft.  9  in.  to  14  ft.  8  in.  apart.  A  better  arrangement  would  have 
been  to  have  these  columns  spaced  about  18  ft.  apart. 

The  author  also  states  that  the  construction  of  the  concrete  work 
in  this  building  during  the  winter  of.  1906-07  was  entirely  successful, 
thus  demonstrating  that  reinforced  concrete  work  may  be  conducted 
during  a  New  York  winter  without  material  interruption.  He  states 
that  this  was  accomplished  by  covering  window  openings  with  canvas, 
using  salamanders,  and  covering  the  fresh  concrete,  as  fast  as  poured, 
with  tarpaulins  or  hay,  or  both.  Now,  while  it  may  be  possible  to  ob- 
tain first-class  concrete  work  in  this  way,  it  is  undoubtedly  a  great 
expense  thus  to  do  the  work,  and  likewise  risky,  as  the  work  may  freeze 
at  any  time,  especially  the  thin  floor  slabs,  from  underneath. 

The  speaker  knows  of  a  number  of  cases  of  collapse  due  to  this 
cause,  and  he  believes  that  the  erection  of  reinforced  concrete  work, 
in  which  there  are  thin  floor  slabs,  undertaken  during  freezing  weather, 
carries  with  it  grave  responsibility  and  uncertainty. 

It  might  be  interesting  to  compare  the  quantity  of  steel  required 
in  reinforcing  the  concrete  work  of  the  McGraw  Building  and  in  that 
of  a  steel  skeleton  structure.  A  complete  steel  frame  structure  for 
the  McGraw  Building,  computed  for  the  same  loads  that  were  used 
by  the  author,  would  weigh  approximately  1  500  tons.  In  the  McGraw 
Building  the  steel  columns  weigh  655  tons,  the  reinforcing  rods  507, 
making  a  total  of  1 162  tons,  equivalent  to  a  saving  in  the  McGraw 
tjTDe  of  340  tons  in  the  steelwork.  This  would  amount  to  about  $21  000, 
assuming  the  price  of  steel  to  be  about  $62  a  ton.  This  difference, 
however,  would  most  likely  be  more  than  offset  in  other  ways  in  the 
steel  skeleton  type,  as  there  would  be  much  less  concrete  required,  and 
the  erection  methods  would  be  less  expensive.  Perhaps  the  author 
made  comparisons  as  to  the  cost  of  these  different  types  of  construc- 
tion ;  if  so,  it  would  be  interesting  to  have  his  figures. 

In  an  experience  covering  more  than  seventeen  years  in  the  con- 
struction of  buildings,  the  speaker  has  had  to  deal  with  practically 
all  kinds  of  materials,  and  has  had  charge  of  a  number  of  reinforced 
concrete  structures.  In  his  opinion,  nothing,  thus  far,  has  been  de- 
vised which  is  comparable  to  the  modern  steel  skeleton  type  of  con- 
struction for  high  buildings,  not  only  for  safety,  but  for  economy, 
speed  in  construction,  and  ability  to  make  the  frame  as  thoroughly 
fire-resisting  as  possible. 

Every  condition  of  loading  can  be  intelligently  taken  care  of  in  a 


Papers.]        DISCUSSION  ON  REINFOllCED  CONCRETE  BUILDING  161 

steel  structure,  the  stress  in  each  member  of  the  frame  being  capable  Mr.  stern. 

of  complete  analysis,  and  the  knowledge  at  hand  to-day  as  to  what 

the  unit  stresses  should  be  has  been  so  thoroughly  tried  out  that  it  is 

safe  to  say  there  is  practically  no  element  of  vincertainty  in  the  design 

of  a  steel  building.     In  a  reinforced  concrete  building,  however,  the 

case  is  otherwise.    The  factor  of  ignorance  is  much  greater.     Most  of 

the  work  is  done  on  the  premises  by  labor  more  or  less  unskilled. 

The  supervision  of  the  work  during  construction  is  of  the  most 
exacting  nature,  and  requires  high  intelligence  and  unremitting 
vigilance.  The  difficulty  of  getting  good  workmanship,  and  of  making 
the  construction  correspond  with  the  plans,  is  very  great,  and,  finally, 
after  the  work  is  finished,  grave  defects  of  workmanship  may  exist  in 
spite  of  all  the  care  exercised. 

L.  J.  Mensch,  M.  Am.  Soc.  C.  E.  (by  letter). — This  paper  has  been  Mr.  Mensch. 
read  with  great  interest  by  the  writer,  and,  while  he  does  not  doubt 
that  the  owners  are  more  than  pleased  with  the  strength  of  the  build- 
ing, he  has  to  take  exception  to  many  statements  made. 

The  structure  cannot  be  called  a  true  reinforced  concrete  build- 
ing, the  columns  being  of  steel,  fire-proofed  by  concrete,  although 
ostentatiously  calculated  as  reinforced  concrete  columns.  Neither  is 
it  the  latest  type  of  reinforced  concrete  building  construction;  it  is, 
in  fact,  the  oldest  type  of  high  building  construction  in  which  rein- 
forced concrete  was  used.  After  the  introduction  of  reinforced  girder 
and  slab  construction,  many  years  elapsed  before  owners  and  archi- 
tects could  be  persuaded  to  allow  the  use  of  reinforced  concrete  columns, 
and,  in  most  cases,  latticed  steel  columns,  fire-proofed  by  concrete, 
were  used.  Of  the  numerous  biiildings  of  this  type,  the  writer  will 
mention  only  the  ten-story  Audit  Office  of  the  French  Government 
at  the  Cours  de  la  Eeine,  Paris.  The  statement,  that  the  McGraw 
Building  is  higher  than  hei'etofore  considered  practicable,  must  be 
contradicted.  The  height  of  the  Ingalls  Building,  in  Cincinnati,  is 
about  220  ft.  above  the  basement,  and  the  height  of  the  Pugh  Power 
Building,  in  the  same  city,  designed  by  the  writer,  is  about  180  ft. 
above  the  basement.  The  latter  is  used  for  the  same  purpose  as  the 
McGraw  Building,  and  has  also  the  same  spacing  of  columns;  and  the 
first  section,  YO  by  335  ft.,  proved  such  a  success  that  the  owner  built 
an  addition  to  it,  making  it  now  about  150  by  335  ft.  and  ten  stories 
high.    Mr.  Douglas  has  shown  clearly  the  waste  of  steel  in  the  columns. 

It  is  true  that  very  few  tests  of  structural  steel  columns  strength- 
ened by  concrete  filling  have  been  made,  and  the  writer  is  pleased  to 
be  able  to  mention  at  least  one  test  which  was  made  by  Dr.  F.  von 
Emperger,  and  published  in  the  July  number  of  Beton  und  Eisen, 
1907.  TwoX-beams,  about  5 J  in.  deep  and  6§  in.  from  center  to  center, 
were  connected  by  eight  flat  irons  2^  by  I  in.  in  a  length  of  13  ft.  This 
column  failed  at  100  000  lb.,  the  I-beams  buckling  separately.     The 


162  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Mensch.  column  was  straightened  out,  and  the  space  between  the  I-beams  filled 
with  concrete,  and  tested  after  six  weeks.  The  composite  column 
failed  at  265  000  lb.  The  iron  section  contained  6.3  sq.  in.,  the  con- 
crete section  contained  31^  sq.  in.,  the  radius  of  gyration  of  the  two 
I-beams  was  2.6  in.,  and,  from  this.  Dr.  von  Emperger  demonstrates 
that  the  carrying  capacity  of  this  cokamn  is  to  be  considered  as  the 
sum  of  the  carrying  capacities  of  the  iron  and  of  the  concrete.  He 
also  mentions  that  the  concrete  completely  separated  from  the  iron, 
and  crushed  into  pieces  from  1^  to  3  ft.  long.  Test  cubes,  cut  from 
such  pieces,  gave  an  ultimate  resistance  of  1 120  lb.  per  sq.  in. 

Although  this  test  may  not  be  entirely  convincing,  it  shows  that, 
in  such  a  column,  the  highest  working  stress  on  the  steel  section  may 
be  allowed  safely,  disregarding  the  concrete,  which  may  be  considered 
as  acting  only  as  a  stiffener.  From  this  it  follows  that  it  would  have 
been  safe  to  reduce  the  size  of  the  columns  of  the  McGraw  Building. 

The  author  is  correct  in  his  statement  that  the  form  work  repre- 
sents the  most  difficult  part  of  reinforced  concrete  construction;  and 
the  success  of  a  contractor,  and  also  the  speed  of  erection,  depend  en- 
tirely on  his  ability  to  organize  his  carpenter  force,  and  to  give  his 
foremen  complete  working  drawings,  omitting  not  the  smallest  detail, 
even  specifying  the  number  and  kind  of  nails;  in  fact,  do  the  work 
on  the  same  basis  as  structural  steelwork.  But  it  is  also  the  duty  of 
the  engineer  to  design  the  building  so  that  the  form  work  is  reduced 
to  a  minimum.  For  example,  lumber  comes  only  in  certain  sizes — a 
so-called  2  by  10-in.  plank,  is  generally  only  11  by  ,9^  in. — and,  if  a 
column  20  in.  square  is  specified,  the  forms  can  only  be  made  by 
ripping  the  planks.  On  the  other  hand,  it  will  be  found  that  a  column 
19|  by  19J  in.  can  be  formed  in  by  using  commercial  lumber,  and  it 
is  absolutely  necessary  that  the  designing  engineer  should  know  the 
commercial  sizes  of  lumber,  as  they  vary  with  different  localities.  The 
same  applies  to  girders  and  beams,  which,  as  a  rule,  cannot  be  obtained 
in  even  dimensions  without  considerable  waste  of  labor  and  material. 

The  use  of  brackets  should  be  carefully  considered.  It  seems  that 
in  most  cases  they  are  adopted  for  good  luck,  with  no  regard  to 
statical  considerations.  The  writer  has  seen  many  brackets,  the 
under  side  of  which  formed  an  angle  of  60°  and  more  with  the  hori- 
zontal, which  were  generally  not  more  than  8  or  12  in.  in  length.  Such 
brackets  add  greatly  to  the  cost,  but  very  little  to  the  strength,  of  the 
structure.  A  little  consideration  would  show  that  it  would  be  cheaper 
to  use  deeper  girders  and  omit  the  brackets.  A  bracket  is  of  im- 
portance only  in  case  the  underside  forms  an  angle  of  not  more  than 
25°  with  the  horizontal. 

The  layout  of  girders  and  beams  should  be  made  as  simple  as  possi- 
ble. The  writer  cannot  say  that  the  distribution  of  girders  and  beams 
in  the  McGraw  Building  is  the  most  economical,  or  the  most  favor- 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING  163 

able,  for  the  form  work.     In  the  Pugh  Building,   the  girders  were  Mr.  jieusch. 

adopted  in  the  direction  of  the  21-ft.  spans,  and  the  beams  in  the 

direction  of  the  14-ft.  spans,  and  were  spaced  about  11  ft.  apart.    This 

reduced  the  number  of  beams,  and  made  the  centering  of  the  girders 

much  simpler,  and  the  little  excess  of  concrete  used  in  the  slabs,  which 

were  reinforced  in  both  directions — a  necessity  in  every  good  design — 

was  more  than  counterbalanced  by  the  saving  in  the  form  work  and 

time.     The  fact  is  that,  although  the  floor  area  of  this  building  was 

more  than  20  000  sq.  ft.,  and  all  the  walls  and  main  partitions  were 

also  of  reinforced  concrete  above  the  third  floor,  the  rate  of  progress 

was  a  story  every  16  days,  with  a  comparatively  small  gang  of  men. 

In  regard  to  the  use  of  a  derrick  tower  with  four  swinging  booms, 
the  writer's  experience  proves  that  the  cost  of  the  handling  of  the 
concrete  and  the  installation  of  such  an  outfit  is  considerably  more 
expensive  than  the  use  of  a  small  elevator  and  concrete  bucket,  which 
empties  into  a  hopper  and  is  hauled  in  two-wheeled  buggies  to  the 
place  where  needed. 

P.  E.  Stevens,  Assoc.  M.  Am.  Soc.  C.  E.  (by  letter).— In  the  de-  Mr.  Stevens, 
sign  of  the  McGraw  Building,  careful  attention  has  been  given  to 
those  details  intended  to  secure  continuity  in  the  beams,  even  though 
the  New  York  Building  Code  does  not  permit  the  designer  to  take 
full  advantage  of  the  increased  strength  resulting  from  such  con- 
tinuity. This  feature  of  the  design  has  been  made  the  subject  of 
some  adverse  criticism,  on  account  of  the  abundant  reinforcement  pro- 
vided. The  writer  believes  that  these  criticisms  are  not  well  founded. 
The  description  and  drawings  of  the  reinforcing  frames  for  the  girders 
show  the  steel  reinforcement  over  the  supports  to  be  the  same  as  that 
at  the  center  of  the  span,  and  this  has  been  regarded  by  some  as  a 
waste  of  material.  The  usual  formulas  for  stresses  in  continuous 
beams  apply  in  the  case  of  concrete  only  when  such  reinforcement  is 
provided.  In  the  derivation  of  such  formulas,  three  conditions  are 
imposed : 

1. — Unyielding    supports,    conforming    to    the    unstrained    out- 
line of  the  beam — usually  styled  supports  all  on  a  level; 
2. — Spans  all  equal; 

3. — Uniform  moment   of  inertia  throughout  the  length   of   the 
beam. 

The  first  of  these  conditions  it  is  impossible  to  fulfill,  the  second 
seldom  prevails,  and  the  third  is  commonly  ignored.  This  is  not  in- 
tended for  cynicism,  but  is  a  simple  statement  of  fact.  Imperfect 
workmanship,  uneven  shrinkage  of  concrete,  and  elasticity  in  the  ma- 
terial, make  any  assumption  of  "supports  on  a  level"  untenable.  This 
alone  is  sufficient  reason  for  placing  small  dependence  on  the  increased 
strength  due  to  continuity,  when  designing  the  beam. 


164  DISCUSSION  ON  KEINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Stevens.  An  example  of  a  very  common  and  erroneous  interpretation  of  the 
third  condition  is  found  in  another  discussion  of  this  subject.  Mr. 
Noble  has  said  that  the  formulas  for  stresses  in  continuous  beams 
apply  only  when  the  bending  moments  are 

"adequately  met  by  moments  of  resistance,  and  then  only  when  the 
Linit  stresses  in  the  material  furnishing  this  amount  of  resistance  are 
the  same  at  the  center  span  and  the  points  of  support." 

This  is  decidedly  at  variance  with  the  third  condition  imposed  by 
the  formulas.  Uniformity  of  stress  is  far  from  identical  with  uni- 
formity of  moment  of  inertia. 

The   amount   of   the  bending  moment   over   the   support    added   to 

V  1- 
that  at  the  center  of  the  span  will  give  a  sum  equal  to    — —    in    an 

infinite  series  of  uniform  spans  uniformly  loaded.  If,  further,  the 
moments  of  inertia  of  the  sections  of  the  beam  are  the  same  through- 
out  its   length,   then,    and   then   only,   the   bending  moment   over   the 

support  is   ^  f__j,  and  that  at  the  center  is    .^  (^-^    )•     ^^^  ^  series 

of  eight  or  more  spans  complying  with  the  three  conditions  before 
mentioned,  and  uniformly  loaded,  the  span  at  the  middle  would  have 
approximately — within  1% — the  above  distribution  of  bending  mo- 
n>ents. 

A  design  which  assumes  some  fraction  of  this  total,    -3—,  as  the 

moment  at  the  support,  and  the  remainder  as  the  moment  at  the  center 
of  the  span,  simply  because  moments  of  resistance  have  arbitrarily 
been  provided  at  those  points  to  resist  such  moments,  cannot  be  justi- 
fied by  any  sound  theory. 

In  order  to  show  how  erroneous  any  conclusions  drawn  from  such 
assumptions  may  be,  the  writer  has  derived  the  correct  moments  at  the 
supports  and  center  of  the  span  corresponding  with  various  ratios  be- 
tween the  moment  of  inertia  for  the  cantilever  portion  and  that  for 
the  suspended  portion  of  the  span. 

The  beam  to  be  considered  will  be  assumed  to  be  one  of  an  infinite 
series  of  uniform  spans,  uniformly  loaded — approximated  by  a  span 
at  the  middle  of  a  series  of  eight  or  more  spans — and  with  supports 
''all  on  a  level."  In  Fig.  9  let  the  following  nomenclature  and  condi- 
tions govern : 

A  E  =  X\\e   undeformed   neutral   axis   of    the    beam,    with 
supports  at  A  and  E ; 
A  B  C  D  E  =  the  deformed  neutral  axis ; 
B  and  D  =  the  points  of  contraflexure ; 

w  =  the  uniform  load  per  unit  length ; 
I  =  the  moment  of  inertia  of  any  section  of  A  B,  or 

D  E; 
K  =  the  moment  of  inertia  of  any  section  of  B  CD. 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING 


165 


All  conditions  are  symmetrical,  therefore  the  elastic  curve  will  be  Mr.  Stevens, 
symmetrical. 

a  =  the  distance,  A  B,  or  D  E; 
h  =  the  distance,  B  I); 

ilf Q  =  the  bending  moment  at  A  or  E j 

.¥^  =  the  bending  moment  at  the  center  of  the  span,   C; 

B  D  will  be  parallel  with  A  E. 

B  F  is  tangent  to  A  B  at  B,  making  the  angle,  v. ,  with  A  E  and 
BI); 

B  O  is  tangent  to  BCD  at   B,  making   the  angle,  fj,  with  B  D; 

n  and  m  are  current  co-ordinates  of  points  in  B  G  D,  referred  to  B ; 

X  and  y  are  current  co-ordinates  of  points  in  A  B,  referred  to  B. 


Fig.  9. 

Since  B  and  D  are  points  of  contraflexure,  and  bending  moments 
at  these  points  are  zero,  the  beam  may  be  regarded  as  made  of  three 
beams:  two  cantilevers,  A  B  and  E  D,  and  a  simple  span,  B  D.  Con- 
sider first  the  span,  B  D,  loaded  with  w  per  unit  of  length.  By  the 
well-known  theory  of  flexure, 

"  ^  WWJv  ^   ~     '"^  +  "'^ ^^^ 

Differentiate  and  obtain  the  first  derivative, 

d  n       r(j  (&^  —  6  &  nr  -\-  4  m^) 

dm  24  E  K  ■   ^ 

Make  m  =  0,  then, 

-"■'^-A '•^" 

Consider  now  the  cantilever,  A  B:  It  is  loaded  with  the  uniform 
load,  »•,  per  unit  length  and  the  end  reaction  from  B  D  at  B.     This 

end  reaction  is  eqiml  to     .-,  .    From  uniform  load: 


166  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Bir.  Stevens.  deflection  =  -J^—  (,r*  —  4  a' x -{- ?,  a*) (4) 

42  E  I  ^ 

IV  b 
and,  from  load,  -^  at  -B; 

w  h 

2 
deflection  =  —^—.(2  cfi  —  3  a^  x  -\-  x^) .  .  . (5) 

and  ?/,  the  sum  of  these  deflections,  is 

2  w  b  (2a^  —  3a^  X  -\-  x^)  +  iv  (a;*  —  4  a^  x  +  3  a^) 

y  = 24^1 -■■■■('^^ 

obtain  the  first  derivative  : 

dy  _2^vh  (Sx^  —  3  a^)  +  w  (4  ar'  —  4  a^) 

d^x  24  EI  ^' ' 

make  x  =  0,  then, 

6  w  b  (^  —  4  w  a^ 

*"^^-    ''= WEI ('^> 

Since  5  is  a  point  of  contraflexure,  v.  =  /3,  and 

tan.   a-  =  tan.  /? (9) 

and,  from  Equations  3  and  8, 

w  b^                6  ui  b  ((-  —  4  tv  a^ 
MEK  ^  2¥WT  ^     ' 

^^  ^ 

whence,  ^  3  =  —  ~      (11) 

6  0  cr  +  4  a"*  I 

The  negative  sign  governing  the  second  term  of  this  equation  is 
due  to  the  fact  that  the  moments  in  the  cantilever  and  those  in  the 
suspended  portion  are  of  opposite  sign.  It  will  be  dropped  hereafter, 
as  it  is  immaterial  to  this  discussion. 

By  assigning  to  a  and  h  consistent  values,  fractions  of  the  total 

span,  A  D  =^  I,  corresponding  values  of    j   may  be  obtained.     Such 
values  have  been  platted  and  a  curve  drawn  through  them  (Fig.  10). 

10    t 

A  curve  has  also   been  drawn  for  the  values  of  M^^  in  terms   of  — -, 

and  one  for  the  ratio,  ^>^,  for  comparison  with     f.     Attention  is  called 
.li^  1 

to  the  values  of  the  various  functions  given  by  these  curves  corre- 

1  A' 

sponding  with  a  =  I  (1  —  .3  ^3)  =  0.211   I.     The  curves  show  y  =  1; 

hence  K  —  J, 

lf„  =  0.00  (^^-) 

which  values  correspond  exactly  with  the  formula  for  beams  with  uni- 
form  moment   of   inertia.     It  will   be  noted   that  the  curves   for 


Papers.]        DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING  167 


Mr.  Stevens. 


100    10 


00      0 


80      8 


IS   10 


60 


o   50  =5 


40       4 


10       1 


^ 

/ 

/ 

/ 

y. 

/i 

/.i 

1 

\l " 

A 
/  \ 

/     \ 

/ 

\\  \ 

5 

/ 

s 

"v.^^\^ 

0.0 


u.'2  0.3 

Value  ol"    -2- 

FiG.  10. 


1G8  DISCUSSION  ON  REINFORCED  CONCRETE  BUILDING        [Papers. 

Mr.  Stevens  l/"  i  c   /^     ►-     •      t         •  i  j?  i 

iiiid  -  -'  intersect  at  a  common  value  oi  0.1  i,  indicating  that  lor  these 
values,  if  the  depths  are  uniform,  the  stresses  at  the  center  and  at 
the  supports  will  be  equal.     If,  as  is  often  done,  y  is  made  equal  to    ^ 

and  a  bending  moment  of  ~  (— s— )  is  provided  for  at  the  supports,  an 
overstress  of  12i%  will  result.  If,  as  is  said  to  be  the  French  practice, 
_(— ^)    is   provided   for   at   the   center   of  the   span,    and    _    (^d^) 

is  provided  for  at  the  supports,  then   —  ^  4,  and  il/,^  =  0.-1:70    ( — —\  . 

Therefore,  if  the  conditions  of  continuity  and  loading  assumed  are 
realized,  the  result  will  be  an  actual  theoretical  stress  21  times  as 
great  as  that  calculated.  This  appears  to  the  writer  to  be  extremely 
bad  designing. 

It  must  be  borne  in  mind  that  the  live  load  to  be  sustained  by  the 
beams  and  columns  of  a  building  is  not  a  uniformly  distributed  one, 
but  a  constantly  varying  set  of  unequal  loads,  sometimes  concentrated 
and  sometimes  distributed  over  varying  areas.  In  a  steel  structure, 
where  each  beam  is  an  independent  span,  a  uniform  live  load,  with 
a  proper  "scaling  down  formtila"  can  be  specified  which  will  enable 
the  designer  to  work  very  close  to  actual  conditions.  In  a  reinforced 
concrete  structure,  where  the  floor  slabs,  the  floor  beams,  and  the 
girders  are  built  as  continuous  beams,  the  stress  in  every  beam,  girder, 
and  column  is  influenced  by  every  live-load  concentration,  whether 
it  is  assumed  so  or  not.  In  such  a  case,  it  is  manifestly  impossible  to 
make  close  calculations,  and  the  most  conservative  estimate  must  be 
placed  upon  the  value  of  continuity  as  a  factor  in  saving  material. 
Also,  if  concentrated  loads  are  to  be  provided  for,  the  reinforcement 
over  the  supports  must  be  made  equal  to  that  at  the  center  of  the 
span,  as  has  been  done  in  the  design  of  the  McGraw  Building. 

In  the  design  of  steel  structures,  experience  has  taught  that  forms 
in  which  the  stresses  in  each  member  can  be  determined  accurately 
are  preferable.  Many  of  the  statically  indeterminate  forms,  such  as 
multiple  intersection  trusses  and  continuous  girders,  have  almost  en- 
tirely disappeared. 

The  workers  in  the  younger  art  of  reinforced  concrete  would  do 
well  to  give  most  respectful  consideration  to  this  idea  of  using  stati- 
cally determinate  forms,  which  has  become  so  general  in  the  design 
of  structures  in  steel — a  material  the  properties  of  which  may  be  far 
more  accurately  determined  or  controlled  than  those  of  concrete. 


Vol.  XXXIV.  FEBRUARY,  1908.  No.  2. 


AMEBIC  AN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED    1852. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


THE  USE  OF  REINFORCED  CONCRETE  IN 
ENGINEERING   STRUCTURES. 

An  Informal  Discussion.* 


By  Messrs.  E.  P.  Goodrich,  Edwin  Thacher,  Sanford  E.  Thompson, 

William  H.  Burr,  T.  Kennard  Thomson,  D,  W.  Krellwitz, 

Guy  B.  Waite,  and  C.  L.  Slocum. 


Presented  at  the  Meeting  of  January  8th,  1908. 


E.  P.  Goodrich,  M.  Am.  Soc.  C.  E. — The  use  of  reinforced  con-  Mr.  Goodrich. 
Crete  in  engineering  structures  has  had  a  phenomenal  development, 
both  as  to  the  amount  built  in  each  succeeding  year  and  as  to  the 
variety  of  applications  made.  Its  field  of  usefulness  is  rapidly 
broadening,  and  its  exploitation  is  believed  to  have  been  overdone  in 
a  few  lines. 

The  theory  concerning  the  mode  of  action  of  the  two  materials 
involved,  is  constantly  undergoing  modification,  making  it  more  per- 
fect through  deduction  from  experiment.  This  is  the  scientific  method 
of  development  of  any  art;  and  in  this  particular  branch  of  the  build- 
ing art,  it  is  believed  that  by  experiment  alone  can  proper  working 
stresses  be  determined,  upon  which  to  base  all  designs  of  structures. 
Such  stresses  should  be  deduced  primarily  from  fatigue  exi^eriments, 
and  not  be  chosen  as  arbitrary  fractions  of  ultimate  strengths.  It 
is  believed,  further,  that  every  careful  designer  should  take  proper 
account  of  the  secondary  stresses  induced  in  structures  like  buildings 
and  arches,  by  the  increasing  permanent  set  caused  by  repeated  loading. 

*  The  discussion  of  this  subject,  for  which  no  formal  paper  was  presented,  is  printed  in 
Proceedings  in  order  that  the  views  expressed  may  be  brought  before  all  members  of 
the  Society  for  further  discussion. 


170  DISCUSSION  ON  THE  USE  OP  EEINFOECED  CONCRETE     [Papers. 

Mr.  Goodrich.  Experimental  research  is  yet  much  needed  along  several  different 
lines  connected  with  this  subject.  More  light  is  desired  as  to  the 
cause  and  cure  of  the  retrogression  in  the  tensile  strength  of  cement 
briquettes,  often  disclosed.  More  extended  compression  tests  are  also 
needed  to  determine  the  presence  and  amount  of  any  retrogression  in 
the  compressive  strength  of  concrete.  When  tests,  published  in  engi- 
neering periodicals,  show,  at  the  end  of  2  years,  values  hardly  greater 
than  those  at  the  end  of  7  days,  it  would  seem  as  if  this  subject  needed 
most  careful  investigation. 

If  possible,  a  cement  of  higher  compressive  strength  should  be  de- 
veloped, especially  for  use  in  concrete  columns.  Perhaps  this  is  im- 
possible with  the  materials  involved,  because  of  their  very  nature,  and 
because  the  strength  of  the  aggregate  has  been  practically  reached; 
but  unless  some  considerable  increase  can  be  secured,  it  would  seem 
as  if  concrete  columns  would  have  to  be  excluded  from  consideration 
in  high  building  design;  that  is,  in  structures  higher  than  perhaps  six 
stories.  In  their  stead,  structural  steel  columns  would  seem  necessary^ 
but  they  should  be  heavily  fire-proofed  and  entirely  filled  with  a  con- 
crete of  cheap  quality.  It  may  seem  excessive  to  some  engineers,  but 
it  is  believed  that  experience  has  shown  the  necessity  of  fully  3  in.  of 
good  concrete  fire-proofing  over  all  extreme  edges  of  such  columns. 

The  best  design  for  the  steelwork  of  columns  of  this  variety,  is 
believed  to  be  of  angles  latticed  or  battened,  of  channels  similarly 
fabricated,  or  perhaps  wide-flanged  I-beams,  or  the  usual  Z-bar  types. 

Columns  of  the  Considere  variety  are  believed  to  be  proper,  if  a 
suitable  relation  exists  between  the  spiral  and  the  longitiidinal  re- 
inforcement, and  if  a  sufficient  quantity  of  each  is  used.  In  such 
columns,  a  lower  limit  should  be  set  on  the  quantity  of  each  kind  of 
reinforcement,  and  an  upper  limit  on  the  size  of  the  opening  between 
the  parts.  It  may  not  be  out  of  place  to  state  that  the  inventor  of 
spiral  reinforcement  himself  uses  spirals  of  very  thick  material  with 
a  comparatively  small  pitch,  and  it  is  believed  that  a  large  majority 
of  the  columns  being  erected  at  the  present  time,  and  considered  of 
high  carrying  capacity,  would  not  disclose  any  appreciable  excess  of 
carrying  power  if  tested  to  failure. 

In  reinforced  concrete  columns,  with  longitudinal  rods  as  the 
principal  reinforcement,  an  upper  limit  should  be  set  on  the  percentage 
which  may  be  allowed.  In  addition  to  the  fact  that  many  laboratory 
tests  show  lower  efficiencies  for  rods  of  large  diameter  in  concrete 
columns,  it  would  seem  as  if  the  use  of  rods  more  than  IJ  in.  in  diam- 
eter, or  aggregating  more  than  5%  of  the  total  area  of  the  column 
were  of  more  than  doubtful  value,  simply  from  the  impossibility  of 
being  certain  that  enough  adhesion  is  developed  to  secure  the  theoreti- 
cal compressive  stress  in  the  steel  itself.     It  might  seem  as  if  more 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  171 

nearly  practical  conditions  would  be  secured,  in  laboratory  tests,  if  m>-- Goodrich, 
all    reinforcement   were   kept    away    from   the    ends    of   the   concrete 
columns  a  distance  equal  to  at  least  one-fourth  of  the  diameter  of  the 
column. 

It  is  believed  that  too  little  attention  is  given  to  the  design  of  the 
footings  under  such  columns,  especially  with  regard  to  a  proper  trans- 
mission of  the  stress  in  the  longitudinal  rods  into  the  foundation 
concrete,  and  that  in  most  work  not  more  than  half  the  proper  number 
of  ties  are  used  to  prevent  buckling  in  the  vertical  rods.  The  dis- 
position of  the  latter,  so  as  to  prevent  easy  depositing  of  the  concrete, 
is  imperative,  and  those  varieties  of  columns  in  which  the  steel  is 
distributed  very  uniformly  through  the  whole  column  section  are  viewed 
with  distrust,  however  superior  they  may  be  considered  from  a 
theoretical  standpoint. 

Great  care  should  also  be  exercised  in  the  design  of  the  beam  and 
girder  reinforcement,  to  prevent  a  congestion  of  steel  in  the  column 
sections  at  the  floor  levels.  Much  ingenuity  can  profitably  be  expended 
in  obviating  this  trouble. 

It  is  well  known  that  the  addition  of  steel  to  increase  the  com- 
pressive strength  of  concrete  columns  is  not  on  the  side  of  economy  of 
first  cost,  but  only  of  economy  of  floor  area  occupied.  It  would  seem, 
therefore,  as  if  the  best  practice  would  be  to  introduce  steel  only  to 
carry  bending  stresses,  and  to  use  a  cement  of  higher  quality  (if  ob- 
tainable), or  a  richer  mixture  of  the  commercial  product,  and  thus 
secure  higher  working  stresses  with  correspondingly  smaller  sections. 

The  subject  of  impervious  concrete  is  of  vital  importance  for  those 
who  are  interested  in  the  construction  of  dams,  reservoirs,  conduits, 
sewers,  and  water  pipes;  of  hardly  less  interest  in  connection  with 
sea  walls,  retaining  walls,  bridge  abutments,  and  building  founda- 
tions; and  even  of  considerable  interest  in  regard  to  the  superstruc- 
tures of  buildings,  arch  bridges,  etc.  Many  experiments  have  been 
made  as  to  the  perviousness  of  different  mixtures  of  different  aggre- 
gates of  different  sizes,  but,  apparently,  something  further  is  neces- 
sary. Great  things  are  claimed  for  the  several  patented  wet  and  dry 
compounds  now  on  the  market,  designed  to  render  concrete  im- 
pervious, and,  until  further  progress  can  be  made  in  this  line,  the  use 
of  the  best  of  these  in  all  concrete  work  is  strongly  recommended. 
Perhaps  an  impervious  cement  will  soon  be  evolved,  produced  either 
by  the  addition  of  one  of  the  present  products  to  the  practically 
finished  cement  product,  somewhat  as  gypsum  is  now  added,  or  by 
some  other  substance  which  the  inventor  may  work  out.  Such  a 
cement  is  greatly  to  be  desired,  if  for  no  other  reason  than  to  prevent 
the  unsightly  discoloration  from  efflorescence  which  now  defaces  al- 
most all  exterior  cement  work.  This  may  be  partially  cured  by  the 
use  of   a  so-called   "non-staining"  cement,  but  all  those  now  on  the 


172  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE      [Papers. 

Mr.  Goodrich,  market  are  unsatisfactory,  either  as  to  effect  or  first  cost.     A  better 
and  cheaper  one  is  essential. 

Although  other  elements  than  moisture  are  involved  in  the  rusting 
of  steel,  the  use  of  an  impervious  cement  would  have  gone  far  toward 
preventing  the  absolute  disintegration  of  the  metal  backing  of  some 
stucco  work  examined  by  the  speaker  on  several  occasions,  and  would 
make  one  feel  much  safer  as  to  the  probable  life  of  some  of  the  re- 
inforcement which  has  been  plainly  visible  on  the  bottoms  of  floor 
slabs  when  the  latter  were  slightly  scratched  with  a  knife  or  other 
sharp  implement.  It  will  not  be  very  long  before  some  of  such  floors 
will  show  signs  of  failure,  especially  those  in  which  wires  of  small 
diameter,  or  sheet  material  of  small  thickness  have  been  used  for  the 
reinforcement. 

Impervious  concrete  is  also  of  vital  importance  in  foundations, 
and  walls  of  reservoirs,  conduits,  etc.,  through  which  water  will  perco- 
late slowly.  It  is  known  that  at  least  one  heavy  retaining  wall  be- 
came honeycombed  to  such  an  extent  that  failure  resulted;  and  in  one 
high  building  on  lower  Broadway,  in  New  York  City,  the  sub-cellar 
walls  have  been  screened,  apparently  to  hide  the  process  which  may 
possibly  be  slowly  producing  a  similar  effect  in  that  structure. 

Again,  the  few  experiments  which  have  been  made  with  regard  to 
electrolysis  of  embedded  steel  in  wet  concrete,  together  with  the 
astonishing  phenomena  observed  in  one  reinforced  concrete  street  car 
barn,  in  which  hot  metal  was  sometimes  encountered  when  a  trolley 
pole  left  the  trolley  wire,  seem  to  be  convincing  evidence  of  the 
necessity  of  using  impervious  concrete  in  all  reinforced  foundations 
which  may  be  in  the  line  of  electric  earth  currents. 

It  would  also  seem  wise  to  use  only  such  concrete  in  reinforced 
concrete  piles,  because  they  are  relatively  slender  members,  and  any 
disintegration  of  either  the  reinforcement  or  the  concrete  in  them 
would  be  of  grave  moment.  As  to  the  general  subject  of  concrete 
piles,  not  enough  is  yet  known.  While  a  strong  difference  of  opinion 
may  exist,  it  would  seem  as  if  fewer  objections  could  be  raised  against 
those  piles  which  are  moulded  in  plain  sight  and  driven  as  is  a  wooden 
pile,  than  against  those  piles  which  are  moulded  in  place.  The  latter 
rarely  are  properly  reinforced,  and  it  is  extremely  likely  that  the 
fresh  concrete  will  be  displaced  before  it  has  properly  set,  by  opera- 
tions in  their  vicinity. 

A  very  heavy  hammer  should  be  used  for  driving  such  piles,  one 
weighing  at  least  as  much  as  the  pile  being  very  important. 

It  is  probable  that  impervious  concrete  will  partially  solve  the 
problems  incident  to  the  use  of  cement  in  sea  water,  whether  the  dis- 
integration caused  by  the  latter  be  chemical  or  mechanical  in  its 
nature.  It  has  already  been  demonstrated,  at  the  New  York  Navy 
Yard,    for    example,    that    a    rich    face    mixture,    rendered   more    im- 


Papers.]      DISCUSSION  ON  THE  USE  OF  llEINFORCED  CONCRETE  173 

pervious   by  careful  surface  treatment  as   soon  as   the  forms   are  re-  Mr.  Goodrich, 
moved,    is    the   best   preventive    of   disintegration.      It   is   hoped    that 
someone,  therefore,  vpill  produce  a  water-proof  cement. 

More  care  than  is  often  taken,  should  be  exacted  with  regard  to 
the  placing  of  reinforcement.  Some  hints  have  been  given  of  the 
evils  incident  to  poorly  designed  columns,  placing  floor  reinforcement 
too  close  to  the  surface,  and  the  dangers  of  electrolysis.  All  these 
may  be  obviated  to  a  great  extent  by  the  exercise  of  care  in  design 
and  execution.  The  compulsory  use  of  reinforcement  fabricated  in 
units,  in  place  of  separate  bars  in  beams  and  girders,  is  advocated. 
It  is  believed  that  the  small  possible  saving  which  is  claimed  for  the 
latter  method  is  more  than  offset  by  the  insurance  that  a  rod  or  two 
will  not  be  accidentally  omitted  from  an  important  member,  or  a  short 
one  thrown  in  to  take  the  place  of  a  lost  longer  one.  Such  cases  have 
been  actually  observed,  even  with  the  most  perfectly  organized  forces, 
and  one  superintendent  of  a  company  which  still  advocates  separate 
bars,  once  said  the  company  could  have  some  of  his  salary  if  they 
would  use  units,  because  of  the  immeasurable  lessening  of  responsi- 
bility on  his  shoulders. 

More  attention  should  be  paid  to  the  subject  of  shear  or  diagonal 
tension  in  reinforced  concrete  beams  and  girders.  The  fact  that  cer- 
tain empirical  systems  have  produced  many  buildings  which  have  not 
collapsed  under  load,  is  no  proof  of  the  adequacy  of  their  reinforce- 
ment in  this  respect.  It  is  believed  that,  in  much  work  now  under 
way,  while  the  factor  of  safety  against  failure  through  tension  or  com- 
pression is  four  or  more,  the  margin  of  safety,  with  regard  to  diagonal 
tension,  is  much  smaller.  The  care  taken  with  this  point  of  design 
by  foreign  engineers  is  far  above  that  common  in  America.  Many 
more  experiments,  covering  various  ages  and  arrangements,  are 
urgently  needed. 

The  ideas  incident  to  the  use  of  discrete  structural  members  are 
not  applicable  to  the  design  of  monolithic  concrete  structures.  In  the 
latter,  the  continuity  of  the  members  should  be  recognized,  and  the 
reinforcement  of  columns,  beams,  girders,  and  floors,  should  be  ar- 
ranged so  as  to  make  the  parts  act  as  rigidly  connected  elements.  The 
desig'n  of  colunms  simply  as  compression  members,  entirely  ignoring 
the  bending  produced  by  unbalanced  loads  on  rigidly  connected  girders, 
is  not  considered  the  best  practice;  and  the  use  of  the  factor,  |,  in  the 
moment  formula  is  an  inheritance  from  the  older  methods  used  in 
timlier  and  steel.  Even  the  use  of  so  much  reinforcement  at  supports 
as  corresponds  with  the  factor,  ^^y,  is  believed  to  be  entirely  inade- 
quate; and  the  speaker  ventures  the  prophecy  that  progressive  failure 
is  taking  place  in  many  structures  designed  with  only  that  quantity 
of  steel  at  the  points  in  question. 

It  is  further  advocated,  that  designs  be  made  so  that  tests  of  se- 


174  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  Goodrich,  curity  can  be  carried  out  on  the  very  day  the  centering  is  allowed 
to  be  removed.  It  is  also  recommended  that  the  tests  be  exacted  on 
those  dates.  Such  specifications,  rigidly  adhered  to,  would  reduce  to 
a  minimum  the  danger  and  number  of  premature  failures. 

Deformed  rods  may  be  better  in  theory,  but  almost  no  practical 
proofs  of  their  superiority  have  been  produced,  as  far  as  known. 
Laboratory  tests  are  hardly  conclusive,  since  many  experiments  on 
beams  actually  seem  to  show  some  kinds  of  rods  to  be  really  detri- 
mental to  the  best  results. 

More  experiments  are  very  desirable  concerning  the  effect  of  the 
proportion  of  water  used  in  the  original  mixture  and  the  effects  of 
continued  and  intermittent  saturation  of  the  concrete,  upon  the  ad- 
hesion between  it  and  smooth  rods  embedded  therein.  Perhaps  the 
use  of  impervious  concrete  would  solve  this  difficulty,  irrespective  of 
the  actual  effects  produced  by  excessive  moisture.  More  fatigue  ex- 
periments, also,  are  essential  to  a  full  knowledge  of  this  subject,  and 
the  very  few  so  far  reported  along  all  lines  are  worthy  of  the  highest 
commendation  and  the  most  careful  study. 

The  compression  experiments  of  this  kind,  in  conjunction  with 
those  carried  to  rupture  on  columns  of  the  Considere  type,  seem  really 
to  show  the  justice  of  allowing  high  stresses  on  such  columns.  As 
long  as  the  elastic  limit  of  the  concrete  is  not  reached,  since  columns 
rcinforced'in  this  way  show  very  large  deformation  before  final  failure 
(thus  reducing  the  danger  of  the  latter),  there  would  seem  to  be  no 
good  reason  for  restricting  the  working  stress  to  the  low  figures  at 
present  usually  exacted  for  plain  concrete  or  longitudinally  reinforced 
columns. 

Nor  do  rods  of  high  elastic  limit  appear  to  be  advantageous,  under 
ordinary  conditions.  Since  all  varieties  of  steel  have  practically  the 
same  modulus  of  elasticity,  and  since  the  first  tension  cracks  in  the 
concrete  appear  at  approximately  the  same  strain  in  all  specimens, 
and  consequently  at  the  same  stress,  irrespective  of  how  much  higher 
the  elastic  limit  may  be,  the  relative  amount  of  the  latter  is  of  no 
importance  provided  it  is  beyond  the  usual  working  stress. 

Perhaps  such  rods  may  be  of  value  in  column  work,  where  high 
stresses  are  used,  and  they  may  be  advantageous  in  the  reinforcement 
of  long  walls  against  shrinkage,  but,  even  in  these  positions,  the  ad- 
vantage is  not  evident.  Reports  as  to  actual  structures  of  the  last 
mentioned  kind,  where  no  cracks  have  appeared,  together  with  the 
amount  of  steel  introduced,  are  greatly  to  be  desired.  It  is  possible 
that  the  distribution  of  the  reinforcement  is  also  influential  to  some 
extent. 

The  character  and  size  of  the  aggregate  does  not  receive  half  the 
attention  it  deserves,  and  the  quantity  of  water  being  used,  especially 
in  the  manufacture  of  much  cement  brick,  concrete  blocks,  and  orna- 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  175 

mental  cement  work,   is  entirely  inadequate.     Very  few  persons  are  M'"  fioodrich. 
interesting  themselves  in  the  artistic  phase  of  the  subject,   and  the 
results  attained  in  most  part  are  still  considered  rather  crude.    There 
are  some  beautiful  exceptions  to  this  statement,  however. 

Experiments  should  be  made  as  to  means  of  securing  more  uni- 
formity of  color  of  stucco,  and  the  application  of  color  to  cement  sur- 
faces should  receive  more  study.  In  Europe  there  are  some  beautiful 
examples  of  such  work.  While  some  progress  has  been  made  in  devis- 
ing effective  and  pleasing  results  in  surface  treatment  of  concrete 
work,  there  is  still  ample  opportunity  for  improvement.  All  are 
familiar  with  the  terazzo  effect  of  good  granolithic  work,  and  most 
have  seen  surfaces  which  have  been  picked,  axed,  hammered,  or  treated 
with  a  sand  blast.  Some  of  the  effects  produced  by  washing,  with 
heavy  scrubbing  while  quite  fresh,  and  of  etching  with  weak  acid  are 
fairly  pleasing,  but  probably  the  use  of  stucco  in  all  its  several  varie- 
ties will  eventually  predominate.  Colored  tile  can  also  be  used,  either 
in  connection  with  stucco  or  in  combination  with  selected  aggregates, 
and  treated  with  water  or  acid  to  bring  out  the  color. 

With  the  wider  use  of  stucco,  the  necessity  of  securing  a  perfect 
bond  between  it  and  the  foundation  material  will  be  more  apparent. 
Several  patented  and  secret  processes  are  now  in  use,  but  none  is  be- 
yond reproach,  and  in  this  there  is  a  wide  field  for  improvement.  When 
eventually  produced,  such  a  bonding  process  should  be  used,  even  be- 
tween parts  of  work  done  on  succeeding  days. 

The  engineer  should  pay  more  attention  to  the  subject  of  forms. 
If  specifications,  hitherto,  had  not  been  so  indefinite  in  regard  to  this 
item,  fewer  premature  failures  would  have  taken  place.  The  practice 
to  be  followed  in  the  erection  of  at  least  one  important  arch,  of  design- 
ing and  specifying  in  detail  all  points  as  to  the  centering,  can  be 
followed  profitably  in  lesser  structures.  With  this  element  of  risk 
removed,  wherein  the  contractor  has  an  opportunity  to  involve  seriously 
the  safety  of  the  work  by  his  faulty  design  and  erection  of  falsework, 
and  with  the  use  of  reinforcement  in  units  designed  by  an  engineer 
of  long  and  wide  experience,  there  is  no  reason  why  reinforced  con- 
crete work,  eventually,  should  not  become  absolutely  safe  and  fairly 
economical.  Only  one  other  point  remains:  the  process  of  manu- 
facture of  concrete  should  be  inspected  as  carefully  as  the  production 
of  structural  steel  and  the  grading  of  timber.  Then  the  ideal  will 
have  been  reached.  Meanwhile,  a  careful  study  of  the  problem  of 
forms  is  exceedingly  profitable,  because,  in  the  cost  of  finished  work, 
that  of  the  labor  and  material  thus  involved  often  exceeds  40%,  and 
sometimes  approaches  75%,  of  the  total  cost;  and,  when  carefully  done, 
it  may  be  reduced  to  25%,  where  conditions  are  favorable.  The  rapid 
deterioration  of  all  form  material,  because  of  wear  and  tear  from  re- 
peated use,  makes  this  item  of  cost  high,  even  when  the  forms  are 


176  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  Goodrich,  used  a  great  many  times.  Doubtless,  metal  will  eventually  be  used  to 
a  great  extent,  although  wood  will  continue  to  be  necessary  for  many 
parts.  Staff  is  being  used  to  excellent  advantage,  even  for  compara- 
tively simple  work,  but  it  is  not  probable  that  its  use  will  ever  be  very 
extensive.  Some  device  which  will  remain  a  permanent  part  of  the 
structure  will  probably  be  used,  because  these  parts  themselves  can  be 
moulded  in  shops  where  few  forms  are  necessary,  and  the  latter  can 
be  used  a  great  number  of  times.  In  a  similar  way,  the  manufacture 
of  structural  members  in  a  factory,  by  machine,  or  in  such  manner 
that  few  forms  are  necessary,  will  also  be  more  widely  developed 
where  conditions  make  it  possible. 

In  the  labor  element,  a  reduction  can  often  be  made  by  handling 
the  forms  in  large  units  by  derricks,  and  many  devices  are  constantly 
being  invented  to  do  away  with  the  costly  work  involved  in  the  use 
of  the  saw,  hammer,  and  nails.  Bolts  and  a  wrench,  and  work  cut  to 
length  in  a  mill,  are  more  nearly  ideal.  In  all  probability,  less  atten- 
tion will  soon  be  given  to  the  finish  of  the  work  as  it  comes  from  the 
forms,  because,  for  most  classes  of  work,  a  better  quality  of  surface 
finish  is  desirable,  and  more  than  enough  money  can  be  saved  by  using 
cruder  forms,  to  cover  the  cost  of  such  surface  treatment. 

Perhaps  it  is  yet  too  early  to  discuss  the  subject  of  standardizing 
the  sizes  of  beams,  percentages  of  reinforcement,  etc.,  but  such  a  step 
will  doubtless  be  taken  just  as  soon  as  the  art  has  outgrown  its  pres- 
ent really  experimental  stage. 

Finally,  a  plea  is  made  for  more  rational  municipal  building  regu- 
lations and  architects'  specifications,  in  the  framing  of  which  engi- 
neers should  have  a  hand.  When  the  designing  engineer  and  the  man 
in  charge  of  the  furnishing  of  materials  and  erection  of  the  work, 
are  distinct  individuals,  better  results  will  be  attained;  and  owner, 
architect,  engineer,  and  contractor  will  then  all  be  striving  for  the 
most  economical  and  artistic  structure  possible. 

Mr.  Thaeher.  Edwin  Thacher,  M.  Am.  Soc.  C.  E.— The  effect  of  sea  water  upon 
Portland  cement  mortar  and  concrete,  and  upon  steel  embedded  there- 
in, is  a  subject  which  has  received  considerable  study  from  American 
and  foreign  engineers  and  chemists,  for  several  years  past ;  but  the  in- 
vestigations thus  far  made  appear  to  have  resulted  in  very  little  positive 
knowledge  on  the  subject.  There  is  considerable  conflict  of  opinion 
between  foreign  experts  themselves,  and  between  foreign  and  American 
experts.  What  is  most  desired  is  to  know  why  certain  works  have 
failed,  and  why  other  works  have  stood  the  tests  of  many  years  with- 
out any  signs  of  decomposition  or  injury.  When  this  is  known  it  will 
be  possible  to  write  specifications  for  future  work  in  which  the  chemi- 
cal composition  of  the  Portland  cement  used,  and  the  mixture,  manipu- 
lation, and  placing  of  the  concrete  shall  be  such  as  will  insure  uni- 
formly  safe   and   satisfactory   results.     According   to   the   best   known 


Papers.]      DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  177 

European  writers  on  the  subject,  the  use  of  Portland  cement  concrete  Mr.  Thacher. 
in  sea  water  is  attended  with  great  risk  of  chemical  decomposition, 
and  it  is  difficult  and  expensive  to  carry  ovit  their  recommendations, 
in  the  way  of  precautions  to  be  observed  to  overcome  partially  the 
risk  of  such  a  result,  and  their  conclusions  do  not  appear  to  be  justi- 
fied by  experience  in  America  during  the  past  twenty  years  or  more. 
M.  Feret  states  that  no  cement  has  yet  been  found  which  will  give 
absolute  security  against  the  decomposing  action  of  sea  water,  that 
sulphuric  acid  is  the  principal  cause  of  decomposition,  that  the  cement 
should  be  low  in  alumina,  and  as  low  as  possible  in  lime,  that  puzzolanic 
material  is  a  valuable  addition  to  the  cement,  that  gypsum  should  be 
used  sparingly,  that  fine  sand  used  in  mixing  is  injurious,  and  finally 
that  the  mortar  must  be  such  as  will  give  a  dense  and  impervious 
concrete. 

Dr.  \V.  Michaelis  also  recommends  a  completely  impervious  mixture, 
but  differs  from  M.  Feret  in  recommending  that  at  least  one-third  of 
the  sand  used  in  mixing  must  be  very  fine.  If  the  whole  body  of  the 
concrete  is  not  impervious,  he  says,  this  impervious  layer  should  sur- 
round the  porous  kernel  on  all  sides,  and  even  underneath.  He  advises 
a  cement  rich  in  silica  and  as  poor  as  possible  in  alumina  and  ferric 
oxide,  also  the  addition  of  puzzolanic  material  to  the  cement. 

M.  Le  Chatelier  considers  that  the  aluminous  compounds  in  Portland 
cement  are  the  direct  cause  of  its  disintegration  in  sea  water,  and  ad- 
vises that  the  alumina  be  replaced  by  oxide  of  iron.  These  foreign 
authorities  do  not  give  the  chemical  composition  of  a  practical  Port- 
land cement,  such  as  they  would  recommend  for  work  in  sea  water, 
but  satisfy  themselves  by  condemning  to  a  greater  or  less  extent  every 
constituent  of  Portland  cement,  except  silica,  and  no  manufacturer 
has  yet  succeeded  in  producing  a  satisfactory  Portland  cement  contain- 
ing this  material  only. 

The  writer  has  communicated  with  quite  a  number  of  American 
engineers  who  have  had  extensive  experience  in  the  use  of  coiaerete 
in  sea  water,  and,  almost  without  exception,  the  results  have  been 
highly  satisfactory,  notwithstanding  the  fact  that  very  little  precau- 
tion has  been  observed  regarding  the  chemical  composition  of  the 
cement,  or  the  impermeability  of  the  mixture;  and  the  damage  sus- 
tained has  been  confined  mostly  to  points  between  high  and  low  water, 
apparently  due  to  mechanical  causes  more  than  to  chemical  decomposi- 
tion. Joseph  E.  Kuhn,  Major,  Corps  of  Engineers,  U.  S.  A.,  Norfolk, 
Va.,  is  of  the  opinion  that  little  apprehension  of  chemical  action  need 
be  felt  when  standard  and  well-proved  brands  of  seasoned  cement  are 
used.  He  mentions  a  sea  wall  built  at  Fort  Monroe,  just  outside  low 
water,  fifteen  years  ago,  of  1  :  4  :  8  concrete,  with  two-man  stone  in- 
corporated. It  has  been  exposed  to  wave  action  from  storms,  iai  which 
the  beach  sand  was  stirred  up,  and  hurled  against  the  wall  with  great 


178  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers.. 

Mr.  Thacher.  force,  also  to  tides  and  heavy  swells  from  steamers.  This  mixture 
would  naturally  give  a  very  porous  concrete,  but  it  is  hard  and  tough, 
and  no  indications  of  chemical  action  or  damage  of  any  kind  are 
noticeable  except  between  high  and  low  water,  where  the  wall  has  in 
places  been  reduced  in  thickness  as  much  as  4  in.  This  face  has  been 
repaired  by  1  :  2J  :  4  concrete.  Major  Kuhn  concludes  that,  by  using 
a  Portland  cement  of  good  quality,  and  a  dense  and  strong  facing  layer 
when  exposed  to  the  action  of  the  water,  concrete-steel  structures  are 
as  safe  in  salt  as  in  fresh  water. 

C.  W.  Staniford,  M.  Am,  Soc,  C.  E.,  Engineer  in  Chief,  Depart- 
ment of  Docks  and  Ferries,  New  York  City,  says : 

"In  the  work  of  constructing  the  bulkhead  or  river  walls  around 
Manhattan,  which  has  been  in  progress  for  the  past  30  years,  and  is 
now  being  continued,  no  extra  precaiitions  are  taken  on  account  of  the 
concrete  being  laid  in  sea  water,  except  the  use  of  first-class  material 
and  careful  work." 

Practically  all  the  river  wall,  from  low  water  up,  has  a  granite 
face,  backed  by  concrete  in  place,  and  heavy  concrete  blocks  set  in 
place  with  derricks  from  low  water  down,  and  the  work  is  in  perfect 
condition,  after,  in  many  cases,  a  period  of  30  years.  This  applies 
also  to  concrete  blocks  laid  above  water  at  points  not  readily  visible, 
and  to  concrete  laid  en  masse  above  low  water  during  the  past  8  years, 
except  in  one  location  where,  between  low  water  and  2^  ft.  above,  the 
concrete  shows  some  signs  of  pitting,  and  slight  disintegration,  which 
indicates  a  wear  occasioned  by  the  extreme  pressure  of  ice  during  the 
long  low-water  slack. 

S.  W.  Hoag,  Jr.,  M.  Am.  Soc.  C.  E.,  Assistant  Engineer,  Depart- 
ment of  Docks  and  Ferries,  says : 

"As  regards  chemical  action,  the  experience  in  New  York  Harbor 
ought  to  be  valuable,  as  our  waters  carry  sewage  probably  not  equalled 
in  any  smaller  city.  If  chemical  action  counts  for  anything,  I  think 
it  would  in  the  harbor  of  New  York  along  the  North  and  East  River 
waterfronts.  I  do  not  think  that  the  possible  deterioration  from  chemi- 
cal action  is  likely  to  amount  to  much,  lanless  the  exposure  is  in  close 
proximity  to  some  chemical  works.  The  above  remarks  are  predicated 
on  first-class  material  and  workmanship." 

A  committee  of  the  Association  of  Railway  Superintendents  of 
Bridges  and  Buildings  made  some  investigation  on  the  subject  of  con- 
crete in  sea  water,  and  some  of  the  replies  to  its  inquiries  are  of  in- 
terest and  may  be  noted  as  follows: 

A,  Wliere  there  is  no  ice,  concrete  made  in  air  with  fresh  water 
and  sunk  in  sea  water,  works  well.  We  would  not  deposit  concrete 
direct  into  sea  water.  Disintegration  more  rapid  than  if  deposited  in 
blocks.  Where  there  is  large  ice  formation,  concrete  between  high  and 
low  water  will  disintegrate  from  ^  to  |  in.  annually.  Stone  facing 
recommended. 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  179 

B.  Mix  dry,  no  water,  and  deposit  through  chutes;   depositing  in  Mr.  Thacher. 
blocks  preferable;  tides  and  frost  have  no  appreciable  effect  on  blocks. 

C.  Concrete  deposited  direct  into  sea  water  gives  perfectly  satis- 
factory results  if  the  materials  and  work  are  right.  The  cement  should 
contain  not  more  than  2%  sulphuric  tri-oxide.  Concrete  should  not 
be  leaner  than  1:2:4.  Stone  facing  preferred  between  high  and  low 
water. 

D.  A  concrete  pier  at  Warren,  E.  I.,  built  about  25  years  ago,  of 
1 :  3  mortar,  is  sound  except  between  high  and  low  tide,  where  it  has 
worn  away  in  places  from  4  to  8  in.,  due  to  ice  and  tide.  Current 
about  8  miles  an  hour. 

The  committee  reports  in  favor  of  depositing  concrete  direct  into 
sea  water.  It  considers  this  method  the  cheapest  and  best,  and  is  of 
the  opinion  that,  with  good  material,  properly  mixed  and  handled,  and 
with  a  granite  face  above  low  water,  it  will  do  good  service. 

Louis  C.  Sabin,  M.  Am.  Soc.  C.  E.,  says: 

''Many  of  the  most  eminent  and  conservative  engineers  consider 
that  most  failures  are  due  to  improper  specifications,  proportions,  and 
manipulation,  rather  than  to  any  defect  in  the  cement." 

William  B.  Mackenzie,  Chief  Engineer,  Intercolonial  Railway  of 
Canada,*  has  used  concrete  in  eight  different  places  in  clear  sea  water, 
and  in  every  case  disintegration  has  taken  place  between  high  and  low 
■tide,  from  J  in.  to  6  in.  in  depth.  The  concrete  was  generally  1:2:4. 
He  learned  that,  where  sea  water  carries  sediment,  the  sediment  pene- 
trates into  the  pores  and  coats  the  surface,  and  no  disintegration  takes 
^lace. 

Martin  Murphy,  Provincial  Government  Engineer,  Nova  Seotia,t 
has  used  concrete  extensively  for  bridge  piers  since  1883.  Some  of  the 
bridges  were  within  the  influence  of  the  turbulent  tides  of  the  Bay  of 
Eundy,  most  of  them  exposed  to  heavy  drift  ice,  and  all  of  tEem  to 
extremes  of  temperature,  yet  but  one  failure  can  be  recorded,  and  that, 
in  his  opinion,  was  due  to  careless  workmanship. 

J.  G.  Theban,  Assoc.  M.  Am.  Soc.  C.  E.,  Engineer  in  Charge  of  the 
Department  of  Bridges,  Borough  of  the  Bronx,  New  York  City,  has 
made  an  interesting  experiment  relating  to  the  preservation  of  steel 
embedded  in  concrete  in  sea  water.  On  August  24th,  1904,  or  some- 
what more  than  three  years  ago,  he  sank  in  Pelham  Bay,  in  20  ft.  of 
vpater,  a  shallow  wooden  box,  in  which  ten  steel  Thacher  bars,  spaced 
at  equal  intervals,  had  been  spiked  to  wooden  cross-pieces.  A  bucket 
of  1:2:4  concrete  was  then  lowered  and  dumped  on  and  around  these 
bars.  After  one  month  the  box  was  raised  and  placed  at  low  tide, 
where  it  was  covered  with  sea  water  twice  every  24  hours.  The  bars 
have  been  removed  from  time  to  time,  and  all  have  been  found  free 

*  Engineering  News,  October  31st,  1907. 
+  Transactions,  Am.  Soc.  C.  E.,  1893. 


180  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE      [Papers. 

Mr.  Thacher.  from  rust.  The  speaker  saw  the  hist  bar  removed  on  January  1st, 
1908,  and  it  and  also  the  spikes  with  which  it  was  fastened  were  free 
from  rust.  Only  a  thin  film  of  grout  at  most  could  find  its  way  under 
the  bars  at  points  where  they  were  in  contact  with  the  wood,  but  no 
rust  could  be  discovered  at  these  points, 
m-.  s.  E.         Sanford  E.  Thompson,  M.  Am.  Soc.  C.  E. — Columns  represent  the 

Thompson.  .  ,•■,•■,■,■  •  ,        r    -i  r  j.i 

most  Vital  part  of  a  building,  since  the  failure  of  one  may  cause  the 
fall  of  the  entire  structure.  The  extreme  variations  in  the  funda- 
mental assumptions  in  different  private  specifications,  and  also  in  city 
ordinances,  make  it  imperative  that  the  subject  should  receive  more 
accurate  and  scientific  treatment.  As  an  illustration  of  the  variety  of 
ideas  as  to  what  constitutes  safety,  the  extremes  may  be  cited  of  cer- 
tain city  ordinances  which  permit  a  load  not  greater  than  350  lb.  per 
sq.  in.  on  the  column,  and  the  value  which  is  sometimes  used  in  private 
practice  of  1  000  lb.  per  sq.  in.  based  on  the  entire  cross-section  of  the 
column  without  appreciable  reinforcement.  The  convincing  argument, 
once  addressed  to  the  speaker  by  a  prominent  architect,  for  the  adop- 
tion of  the  latter  value  in  an  important  structure  was  that  buildings 
in  the  Middle  West  had  been  designed  and  constructed  with  this  unit 
compressive  stress  and  were  still  standing. 

The  owners  of  a  building  frequently  bring  great  pressure  to  bear 
upon  the  designer  to  reduce  the  size  of  the  columns  in  the  lower 
stories.  This  is  not  to  be  wondered  at  when  it  is  considered  that 
their  dimensions  may  be  30  or  36  in.  square,  and  thus  require  an  ap- 
preciable amount  of  floor  space. 

It  is  well  to  recognize  at  the  start  that  reinforced  concrete  columns, 
of  a  section  which  will  compare  favorably  with  steel,  cannot  yet  be 
safely  and  economically  constructed.  A  design  after  the  principles 
followed  by  Professor  Burr  .in  the  McGraw  Building  perhaps  ap- 
proaches a  minimum  section  as  closely  as  is  possible,  but,  even  here, 
only  a  low  unit  stress  can  be  allowed  on  the  steel  without  over-com- 
pressing the  concrete.  It  may  be  laid  down  as  a  general  principle  that, 
not  only  is  it  cheaper  to  resist  compressive  stress  with  concrete  than 
with  steel,  but  also  that  concrete  is  cheaper  than  any  combination 
which  may  be  made  of  steel  and  concrete. 

In  order  to  reduce  the  size  of  concrete  columns,  four  distinct 
methods  have  been  used : 

(1). — Rich  proportions, 

(2). — Vertical  reinforcing  steel, 

(3). — Structural  steel  reinforcement, 

(4). — Hooping  or  banding. 
The  use  of  a  very  rich  mixture  has  much  to   commend  it.     The 
ultimate  strength,  by  using  a  1  :  1  mortar,  may  reach  5  000  lb.  per  sq. 
in.,*  and  the  modulus  of  elasticity  will  also  be  so  high  that  the  defor- 
mation will  be  slight. 

*  '-Tests  of  Metals,"  U.  S.  A.,  1904,  p.  386. 


Papois.]      DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  181 

The  introduction  of  vertical  steel  rods  is  indicated  by  the  majority   Mr.  s.  E, 

.  11.  Thompson, 

of  tests*  to  be  a  satisfactory  manner  oi  increasing  the  strength,  but 

the  low  stress  which  can  be  taken  by  the  steel  without  permitting  too 
.  great  deformation  of  the  concrete,  makes  this  an  expensive  method, 
and  the  percentage  of  steel  is  limited,  not  only  by  economical  con- 
siderations, but  also  because  of  the  difficulty,  especially  when  deformed 
rods  are  used,  of  placing  the  concrete  around  them  properly. 

The  use  of  structural-steel  shapes  for  reinforcement  has  already 
been  so  fully  considered  in  previous  discussions  that  no  further  men- 
tion need  be  made  of  it  here. 

Hooping  or  banding,  first  introduced  by  Considere  in  France,  per- 
haps more  than  any  other  method  of  reinforcement,  has  caught  the 
popular  eye,  with  a  resulting  tendency  to  great  extremes  of  loading. 
For  this  reason,  it  behooves  engineers  to  examine  very  carefully  the 
underlying  principles  involved  in  this  method  of  reinforcement  and 
the  results  of  experiments  thus  far  made. 

To  illustrate  the  position  taken  by  many  conservative  engineers 
on  the  subject  of  hooped  columns,  it  may  be  worth  while  to  study  for 
a  moment  the  real  action  which  takes  place  under  loading,  as  shown 
both  by  theory  and  tests. 

When  a  load  is  placed  upon  the  top  of  any  column,  it  causes  verti- 
cal compression  or  deformation  with,  at  the  same  time,  a  lateral  ex- 
pansion. The  lateral  expansion  in  concrete  columns,  as  shown  by  tests 
upon  plain  and  upon  reinforced  columns  by  Mr.  J.  E.  Howard  at  the 
Watertown  Arsenal,t  and  by  A.  N.  Talbot,  M.  Am.  Soc.  C.  E.,  at  the 
University  of  Illinois,:};  is  at  first  very  small.  Any  stress  produced  in 
the  steel  hooping  must  be  proportional  to  its  deformation  or  stretch- 
ing; hence,  with  small  lateral  expansion  of  the  concrete,  there  can  be 
but  slight  stress  in  the  hoops.  For  this  reason,  and  also  because  of 
the  initial  shrinkage  of  the  concrete,  which  the  lateral  expansion  must 
first  overcome,  scarcely  any  stress  or  pull  comes  upon  the  hoops  until 
the  concrete  within  them  has  reached  a  loading  equal  to  the  breaking 
load  in  plain  concrete.  As  this  load  is  approached,  the  modulus  of 
elasticity  of  the  concrete  becomes  very  much  lower,  and  consequently 
both  the  vertical  and  lateral  deformations  become  much  greater.  Then, 
and  not  until  then,  does  the  lateral  pressure  begin  to  act  appreciably 
upon  the  hoops.  In  other  words,  up  to  the  very  crushing  strength  of 
plain  concrete,  the  value  of  the  hooping  is  actually  negligible.  From 
then  on,  the  reinforcement  takes  practically  all  the  load,  and  a  high 
ultimate  strength  may  be  attained,  although  coincident  with  great 
shortening  of  the  column. 

It  is  evident  that,  if  concrete  is  confined  in  a  tube,  advantage  can 
be  taken  of  the  added  strength  due  to  the  tube.     On  the  other  hand, 

*  "Tests  of  Metals,"  U.  S.  A..  1904,  p.  386;  1905,  p.  377. 

t  '-Tests  of  Metals,"  U.  S.  A.,  190.5,  pp.  293-336. 

t  Proceedings,  American  Society  for  Testing  Materials,  Vol.  VII,  1907.  p.  3H2. 


182  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE      [Papers, 

Thonipson.  ^^  lioops  are  very  far  apart,  it  is  evident  that  the  concrete,  when  it 
reaches  a  stress  equal  to  the  strength  of  plain  concrete,  will  be  thrust 
out  between  the  hoops.  Professor  Talbot's  tests,*  using  a  gradually 
increasing  load,  indicate  that,  with  ordinary  spacing  (the  effect  of 
different  hoop  spacing  is  not  definitely  discussed  in  the  advance  report 
of  the  tests  thus  far  made),  the  hoops  will  effectually  restrain  the  con- 
crete within  them.  The  effect  of  repeated  and  continued  loading  was 
not  investigated  by  him. 

Even  with  the  concrete  restrained  within  the  hoops,  the  shell  of 
concrete  outside  of  them,  which  is  necessary  for  fire-proofing  and  for 
the  protection  of  the  steel,  begins  to  crack  and  peel  at  about  the  same 
load  as  that  which  will  cause  complete  failure  in  unreinforced  con- 
crete. Professor  Talbot,  in  fact,  states  as  a  general  proposition  that: 
"Cracking  and  peeling  of  the  concrete  appear  at  loads  corresponding 
to  the  ultimate  strength  of  the  concrete." 

This  applies  to  hoops  held  rigidly.  If  the  hooping  is  in  short  spiral 
sections,  with  the  ends  of  the  wire  or  rods  simply  lapped  or  insecurely 
fastened  together,  it  follows,  inevitably,  that  the  spiral  must  give  way 
and  unwind  as  soon  as  it  is  exposed  by  the  stripping  of  the  concrete 
from  the  steel.  Consequently,  the  breaking  strength  of  a  column  hooped 
in  this  way  will  only  be  effectively  equal  to  that  of  an  unreinforced 
column. 

The  modulus  of  elasticity  of  the  concrete  within  any  hooping,  after 
the  point  of  exterior  cracking  is  reached,  drops  very  rapidly,  reaching, 
in  the  two  diagrams  shown  in  Professor  Talbot's  paper,  less  than 
300  000  lb.  per  sq.  in.,  even  at  2  000  lb.  per  sq.  in.  stress  in  the  column, 
the  deformation  becoming  so  great,  in  fact,  that  any  vertical  rein- 
forcing steel,  unless  in  such  quantity  as  to  take  the  full  load,  would 
pass  its  elastic  limit  soon  after  the  point  of  first  crack,t  and  by  its 
buckling  increase  the  surface  peeling.  Furthermore,  from  the  ap- 
pearance of  the  deformation  curve,  the  concrete  itself  would  seem  to 
be  in  somewhat  the  same  condition  as  is  steel  after  it  has  passed  its 
elastic  limit. 

When  it  is  considered  that  the  usual  practice  in  concrete  column 
design  takes  no  definite  account  of  eccentric  loading,  or  of  bending 
caused  by  expansion  and  contraction  of  floor  and  wall  areas,  and  that 
inferior  spots  may  occur  in  any  concrete,  through  careless  mixing  or 
placing,  it  appears  that  the  greatest  care  should  be  exercised  in  fixing 
the  unit  stresses  in  hooped  columns. 

Tentative  conclusions  with  regard  to  hooped  column  design  at  the 
present  stage  of  tests  may  be  summarized  as  follows: 

(1). — Hooping,  if  properly  applied,  increases  the  ultimate  break- 
ing strength  under  a  single  loading  to  double  or  treble  the  breaking 
strength  of  a  plain  column. 

*  Proceedings,  American  Society  for  Testing  Materials,  Vol.  VII,  1907,  p.  383. 
t  See  also  Mr.  Howard's  tests,  in  "  Tests  of  Metals,"  U.  S.  A. 


Papers.]     DISCUSSION  ON  THE  USE  OF  EEINFORCED  CONCRETE  183 

(2). — The  surface  of  concrete  outside  of  the  hooping  will  begin  to  q.^Q;^fp^^n 
crack  at  a  loading  corresponding  to  the  breaking  load  of  an  unhooped 
concrete  column. 

(3). — Hooping,  if  not  continuous  or  rigid,  will  peel  off  with  sur- 
face concrete,  so  that  the  effective  strength  of  the  column  will  be  no 
greater  than  a  similar  one  of  plain  concrete. 

(4). — The  total  vertical  deformation  of  a  hooped  column  is  so 
great  at  the  period  of  first  external  crack  that  any  vertical  steel,  unless 
designed  to  carry  the  entire  load,  is  stressed  beyond  its  safe  strength. 

(5). — The  ultimate  breaking  strength  of  a  hooped  column  is  no 
measure  of  its  safe  strength,  and  formulas  based  on  the  ultimate 
strength  are  useless. 

(6). — With  the  present  knowledge,  based  on  tests  in  America  and 
abroad,  the  safe  load  allowed  on  hooped  cokimns  should  be  but  slightly, ' 
if  any,  greater  than  on  similar  columns  without  hooping. 

In  spite  of  the  favorable  reports  which  have  resulted  from  the 
European  experiments  upon  hooped  concrete,  it  seems  impossible  to 
ignore  the  additional  facts  brought  ovit  by  American  tests.  Before  the 
hooping  acts,  the  concrete  has  begun  to  crush,  and  any  structural  ma- 
terial which  has  begun  to  crush  is  unsafe. 

William  H.  Burr,  M.  Am.  See.  C.  E. — Statements  made  in  the  Mr.  Burr, 
course  of  this  discussion  appear  to  indicate  that,  in  such  a  general  treat- 
ment of  the  entire  concrete-steel  question  as  this,  some  featiires  at 
least  of  the  use  of  concrete-steel  should  receive  a  more  careful  con- 
sideration than  would  otherwise  seem  necessary,  in  view  of  recent 
successful  constructions. 

Caution  has  been  urged  against  using  a  unit  working  strees  in  the 
concrete-steel  combination  exceeding  one-tenth  of  the  ultimate  resist- 
ance of  plain  concrete,  such  caution  being  based  upon  some  of  the 
results  obtained  in  the  tests  of  12-in.  cubes  of  1:2:4  concrete  at  the 
Watertown  Arsenal.  In  the  consideration  of  experimental  results 
attained  by  testing  concrete  cubes,  it  is  of  the  utmost  importance  to 
know  completely  all  the  circumstances  of  such  tests,  including  the 
preliminary  tests  of  the  cement  used  and  the  gradations  of  the  sand 
and  gravel  or  broken  stone  aggregate.  If  a  1:2:4  concrete  should 
be  mixed  relatively  dry,  and  allowed  to  set  in  air  and  remain  in  a  dry 
building,  from  the  time  of  its  mixture  until  testing,  the  results  at  the 
end  of  any  usual  test  period  might  and  probably  would  be  quite  dif- 
ferent from  those  found  at  the  end  of  the  same  period  with  a  com- 
paratively wet  mixture  kept  constantly  moist  by  sprinkling  for  a  month 
or  longer  subsequent  to  mixing.  Other  conditions  equally  productive 
of  varying  results  can  be  named,  besides  the  quality  of  the  cement. 

As  a  matter  of  fact,  there  are  numerous  tests  of  12-in.  cubes  of 
1:2:4  concrete  in  the  records  of  the  Watertown  Arsenal  which  show 
an  ultimate  compressive  resistance  of  from  3  000  to  3  600  lb.  per  sq.  in., 


184  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  Burr,  and  even  more,  at  the  end  of  three  months,  with  increasing  resistances 
for  longer  periods.  It  is  a  conservative  statement  to  say  that  well- 
balanced  1:2:4  concrete,  made  with  a  good  quality  of  Portland  cement, 
may  give  from  2  700  to  3  000  lb.  per  sq.  in.,  at  the  end  of  three  months, 
with  ultimate  resistance  continually  increasing  with  age.  Such  con- 
crete may  properly  and  safely  be  expected  to  reach  ultimate  resistances 
2>i  from  4  000  to  4  500  lb.  per  sq.  in.  at  the  end  of  a  year,  results  which 
are  jvistified  by  extended  experience  both  in  America  and  in  Europe. 
It  is  difficult  to  assign  any  satisfactory  reason  for  the  use  of  a  work- 
ing stress  as  low  as  one-tenth  the  ultimate  resistance  of  concrete.  It 
is  true  that  there  are  occasional  cases  of  retrogression,  but,  with  the 
high  grade  of  Portland  cement  available  from  the  most  reputable  pro- 
ducers both  in  America  and  abroad,  it  is  reasonable  to  state  that,  with 
the  usual  engineering  inspection  to  which  the  best  classes  of  public 
work  are  now  subjected,  cement  with  retrogressive  qualities  may  con- 
fidently be  excluded.  No  engineer  at  the  present  time  need  apprehend 
sensible  difficulty  in  securing  Portland  cement  the  resistance  or 
strength  of  which  will  go  on  increasing  indefinitely,  and,  having 
reached  its  maximum,  hold  it.  Under  such  conditions,  a  working  re- 
sistance or  permissible  intensity  of  compression  in  concrete  of  one-fifth 
to  one-sixth  of  its  ultimate,  certainly  affords  all  margin  of  safety  re- 
quired for  engineering  works  of  the  best  class.  Indeed,  probably  a 
somewhat  higher  working  stress  than  that  is  justified  in  large  struc- 
tures of  reinforced  concrete,  especially  where  the  reinforcement  is  of 
such  a  character  as  to  give  material  lateral  support  to  the  concrete. 
This  subject  is  illustrated  effectively  by  the  report  of  a  French  Govern- 
ment Commission  bearing  upon  the  use  of  reinforced  concrete  in 
France.  In  that  report  the  limit  of  compressive  stresses  allowed  in 
reinforced  concrete  is  two-sevenths  of  the  ultimate  crushing  resistance 
of  the  same  concrete  as  determined  by  tests  of  plain  cubes  at  the  age 
of  90  days,  with  the  further  provision  that  this  two-sevenths  may  be 
increased  to  three-fifths  if  the  longitudinal  and  transverse  reinforce- 
ments comply  with  certain  prescribed  conditions.  This  French  pro- 
vision would  yield  a  safe  working  stress  with  first-class  reinforced  con- 
crete work  but  little  if  any  under  900  lb.  per  sq.  in.  The  Bureau  of 
Buildings  of  the  Borough  of  Manhattan,,  New  York  City,  therefore, 
has  taken  a  safe  and  satisfactory  course  in  allowing  750  lb.  per  sq.  in. 
in  such  reinforced  concrete  work  as  the  Thirty-ninth  Street  Building 
in  the  City  of  New  York.  In  fact,  this  latter  working  resistance  is 
conservative  for  the  best  class  of  reinforced  concrete  work  of  the 
present  time. 

The  apprehension  regarding  the  reliability  and  durability  of  rein- 
forced concrete  work  as  shown  by  timorous  expressions  reminds  one 
strongly  of  the  attitude  which  some  engineers  and  others  used  to  take 
toward  structural  steel  when  it  first  came  into  use,  twenty-five  or  more 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFOKCED  CONCRETE  185 

years  ago.  It  is  remarkable,  when  one  reflects  that  structural  steel  is  Mr.  Bun-, 
practically  the  only  structural  metal  which  we  now  possess,  that  at  the 
period  to  which  allusion  is  made  it  was  frequently  argued  out  of  any 
future  possibility  of  use,  as  compared  with  such  a  reliable  material  as 
wrought  iron,  in  consequence  of  the  erratic  behavior  which  some  struc- 
tural steel  members  exhibited  at  that  time.  Fine  cracks,  started  at  a 
punched  rivet  hole  or  sheared  edge,  would  sometimes  extend  far  enough 
to  destroy  the  reliable  carrying  power  of  a  channel  or  angle  or  other 
member.  Such  disclosures,  with  other  erratic  experiences,  were  sources 
of  keen  apprehension  to  many;  others,  however,  believed  them  to  be 
merely  passing  phases  of  difficulty,  which  attend  the  introduction  of 
all  new  materials  and  processes,  and  careful  study,  with  intelligent 
shop  manipulations,  has  shown  them  to  be  such.  Experience,  of 
course,  has  more  than  justified  the  advocates  of  structural  steel,  and 
that  metal  has  now  proved  to  be,  not  only  reliable,  but  by  far  the  best 
structural  material  ever  yet  made  available  to  the  engineer  for  a  wide 
range  of  purposes;  indeed,  wrought  iron  is  no  longer  available  for 
structural  purposes,  nor  has  it  been  for  a  number  of  years. 

Reinforced  concrete  is  passing  through  a  similar  phase.  It  is 
admirably  adapted  to  a  great  range  of  structural  purposes.  Much  has 
already  been  learned  in  regard  to  it,  but  extending  experience  will  dis- 
close a  widening  fund  of  information  of  value  to  the  engineer  in  its 
intelligent  application.  As  a  matter  of  fact,  more  is  actually  known 
about  the  carrying  capacity  or  the  ultimate  resistance  of  concrete- 
steel  members  than  about  the  carrying  capacity  of  steel  columns,  as 
determined  by  actual  tests.  There  has  already  been  accumulated  a 
great  mass  of  well-considered  and  well-digested  experimental  data  re- 
garding the  design  and  construction  of  both  concrete-steel  beams  and 
columns,  although  there  is  need  of  many  additional  tests  of  some  of 
the  latest  and  best  forms  of  concrete-steel  columns.  On  the  other 
hand,  there  are  almost  no  tests  of  full-sized  steel-built  coUimns,  made 
in  such  a  way  as  to  disclose  some  of  the  most  important  fundamental 
principles  of  design.  In  the  present  condition  of  actual  tests  of  the 
two  classes  of  members,  it  is  reasonable  to  believe  that  there  may  be 
at  least  as  much  confidence  attached  to  the  computed  ultimate  carrying 
capacity  of  both  reinforced  concrete  beams  and  columns  as  now  built 
imder  the  best  design  as  can  be  attached  to  the  computed  ultimate 
carrying  capacity  of  steel  columns.  Engineers  have  been  so  accus- 
tomed to  design  and  construct  built-steel  columns  in  their  every-day 
work  that  few  ever  reflect  on  the  paucity,  or  even  absence,  of  experi- 
mental data  on  which  to  base  a  rational  and  competent  design  of  such 
members. 

All  that  reinforced  concrete  construction  needs  for  reliable  results 
is  good  cement,  good  inspection,  and  intelligent  design,  which,  up  to  the 
present  time,  it  has  not  always  had.     It  is  one  of  the  most  useful  build- 


186  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  Burr,  jjjg  inaterials  which  the  engineer  has  yet  had  at  his  command,  but  it 
must  be  dealt  with  in  a  manner  suitable  to  any  first-class  engineering 
work.  There  must  be  rational  design,  intelligent  and  effective  hand- 
ling, and  good  inspection,  precisely  as  with  structural  steel;  and,  under 
such  conditions,  reliable  and  durable  results  may  confidently  be  ex- 
pected. 

Mr.  T.  K.  X.  Kennard  THOMSON,  M.  Am.  Soc.  C.  E.  (by  letter). — Reinforced 
'  concrete,  like  all  other  good  things,  should  be  protected  from  its 
friends.  Many  young  men,  having  very  little  knowledge  of  steel  or  con- 
crete, have  formed  companies  to  build  reinforced  concrete  structures, 
and  one  of  the  first  things  with  which  they  come  in  contact  is  the 
fact  that  to  obtain  a  contract  they  must  bid  low,  another  is  the  neces- 
sity of  showing  the  advantages  of  reinforced  concrete  over  structural 
steel,  and,  as  the  question  of  cost  is  the  one  that  appeals  most  forcibly 
to  the  majority  of  purchasers,  they  try  to  design  their  structure  so 
that  the  cost  will  be  as  low  as,  or  not  much  higher  than,  plain  steel. 
One  of  the  methods  of  doing  this  is  to  use  fiber  strains  which  are 
higher  than  a  good  bridge  or  building  designer  is  accustomed  to  allow. 
Many  who  design  reinforced  concrete  strain  their  steel  bars  up  to 
20  000  or  22  000  lb.  per  sq.  in. — strains  which  bridge  engineers  have 
countenanced  only  for  very  long  spans,  that  is,  those  where  the  dead 
loads  are  large  compared  with  the  live  load.  The  recent  collapse  at 
Quebec,  where  it  was  intended  to  allow  a  possible  strain  of  24  000  lb., 
and  where,  owing  to  faulty  detailing,'  the  structure  failed  at  about 
18  000  lb.,  has  made  many  doubt  the  wisdom  of  allowing  such  high 
combinations  of  strains  (even  if  only  possible),  which  are  hardly  likely 
to  occur  on  any  span. 

It  is  practically  impossible  to  ascertain  the  exact  elastic  limit  of 
the  built-up  members  of  a  bridge — due  to  imperfections  of  workman- 
ship, material,  etc.,  etc.,  and  therefore  it  is  decidedly  unsafe  to  ap- 
proach too  close  to  the  elastic  limit,  in  estimating  the  stresses,  or  to 
assume  that  the  elastic  limit  of  the  test  bar  is  the  elastic  limit  of  the 
full-sized  member.  There  is  no  reason  for  allowing  higher  fiber  strains 
in  reinforced  concrete  than  in  plain  steel,  as  there  are  many  elements 
of  uncertainty  in  the  former  which  do  not  occur  in  the  latter,  because 
far  more  care  is  required  in  the  field  work  and  inspection  of  concrete. 
One  source  of  danger,  "dry  concrete,"  is  rapidly  disappearing,  for 
dry  concrete  practically  required  an  inspector  for  each  laborer,  in 
order  to  ensure  proper  ramming,  whereas  wet  concrete  will  almost 
ram  itself — the  only  danger  being  the  risk  of  letting  the  water  escape, 
thus  carrying  the  cement  with  it.  A  4-in.  reinforced  concrete  wall  in 
New  York  City  was  recently  removed,  when  it  was  found  that  there 
was  no  bond  between  the  steel  and  the  concrete.  Not  knowing  the 
conditions  under  which  the  wall  was  built,  it  can  only  be  assumed 
that  the  concrete  must  have  been  put  in  too  dry. 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  187 

After  the  design  for  a  reinforced  concrete  structure  has  been  made,    -^i''-  T.  k. 

.  ,  .  ,  1         TT  <•     1       Thomson. 

the  three  most  important  considerations  are  the  proper  handhng  of  the 
material,  protection  from  rust,  and — more  important  still — protection 
from  electrolysis. 

In  ordinary  structures  where  large  masses  of  concrete  are  used, 
buckets  containing  2  cu.  yd.  can  be  dumped  in  place,  and,  if  wet,  re- 
quire almost  no  handling,  but,  in  most  reinforced  concrete  structures 
comparatively  little  material  is  used,  and  the  utmost  care  is  required 
in  handling.  In  many  cases  the  extra  sum  paid  for  labor  plus  the 
reinforcement  makes  the  work  cost  as  much  as,  or  more  than,  a  good 
plain  concrete  structure  containing  twice  as  much  concrete,  in  which 
case  it  is  better  to  put  one's  money  into  the  material  rather  than  into 
the  labor. 

Much  difference  of  opinion  exists  as  to  whether  or  not  concrete 
can  be  made  water-tight.  The  writer's  experience  has  been  that  it 
can  be,  but  may  not  always  be,  owing  to  carelessness,  and  that  the  mix- 
ture should  always  be  rich,  that  is  1  part  of  cement  to  2  parts  of 
sand,  with  as  much  stone  as  can  be  covered. 

The  writer  has  seen  24-in.  I-beams,  which  had  been  buried  in  con- 
crete under  the  city  streets  for  five  or  six  years,  taken  out  cleaner  than 
they  were  put  in,  and  in  many  places  showing  the  original  blue  shop 
scale — no  paint  or  oil  having  been  used.  In  a  few  isolated  places, 
however,  these  beams  were  pitted  with  rust,  showing  where  the  water 
had  found  its  way  to  them.  It  is  well  known  that  paint  and  oil  inter- 
fere with  the  bond  between  steel  and  concrete.  Steel  caissons  and 
coffer-dams  have  been  sunk  in  quicksand  in  New  York  City,  which, 
when  exposed  some  seven  years  later,  showed  not  the  slightest  evidence 
of  rust. 

The  writer  has  removed  old  steel  and  cast-iron  columns,  which  had 
been  bedded  in  concrete  and  brickwork  for  years,  which  showed  abso- 
lutely no  sign  of  rust.  Therefore,  in  large  buildings,  carefully  con- 
structed, it  would  seem  that  there  is  almost  no  danger  of  rust,  but  it 
is  doubtful  if  this  is  true  of  reinforced  concrete  bridges,  where  thin 
layers  of  concrete  are  used,  for  it  has  been  found  very  difficult  to  put 
in  a  roadway  floor  which  will  not  allow  any  water  to  percolate  through. 

The  danger  from  electrolysis  is  probably  very  much  greater  than 
from  rust,  and  its  action  is  m.ore  rapid.  There  have  been  cases  in 
New  York  City  where  a  certain  amount  of  current  has  been  grounded 
through  the  steel  in  foundations  buried  in  concrete,  and  the  steel  has 
been  absolutely  destroyed.  For  foundations,  it  would  seem  to  be 
safer,  in  many  cases,  to  rely  on  mass  concrete  rather  than  on  thin 
slabs  of  reinforced  concrete,  which  cost  almost  as  much  in  the  first 
place.  Of  course,  in  cases  where  water  can  reach  the  embedded  steel 
and  carry  an  electric  current  with  it,  the  danger  is  very  great,  and 
very  certain  in  its  action. 


188  DISCUSSION  ON  THE  USE  OF  REINEOKCED  CONCRETE     [Papers. 

Mr.  T.  K         It  is  probably  true  that  steel  in  reinforced  concrete  is  much  less 
Thomson.  |-j^gjy  ^^  ^.^g^  ^j^^^^  -^^  ^  g^gg^  structure  covered  with  the  best  paint, 
but  the  latter  can  be  inspected  and  the  former  cannot. 

In  short,  the  best  friends  of  reinforced  concrete  should  restrict  its 
use  to  its  legitimate  spheres,  which  are  many. 
Mr.  Kreiiwitz.        D.  W.  Krellwitz,  Jun.  Am.  Soc.  C.  E.  (by  letter). — Probably  the 
most  novel  form  in  which  reinforced  concrete  has  been  used  is  in 
transmission-line  structures. 

One  case  is  the  12-mile  transmission,  for  many  thousand  horse- 
power at  high  voltage,  from  Decew  Falls  to  Welland,  Ont.,  Canada, 
for  which  a  line  with  reinforced  concrete  towers  was  completed  in 
1907.  Another  example  is  the  line  of  towers*  carrying  transmission 
circuits  of  high  voltage  to  St.  Catharines,  Ont.  These  towers  are  at 
present  the  highest  monoliths  that  have  ever  been  erected,  being  con- 
siderably more  than  twice  the  height  of  any  of  the  famous  Cleopatra 
needles. 

For  the  elevations  above  ground  at  which  it  is  common  to  support 
the  conductors  of  transmission  lines  (from  25  to  45  ft.),  a  reinforced 
concrete  tower,  in  various  parts  of  the  United  States  and  Canada,  will 
cost  from  one  to  five  times  as  much  as  a  wooden  pole.  It  follows  at 
once  from  this  fact  that  there  must  be  cogent  reasons,  apart  from 
the  matter  of  first  cost,  if  the  substitution  of  reinforced  concrete  towers 
for  wooden  poles  on  transmission  lines  is  to  be  justified  on  economical 
grounds.  The  electric  transmission  of  energy  from  distant  water- 
powers  to  important  centers  of  population  has  grown  from  the  most 
humble  beginnings  to  the  delivery  of  hundreds  of  thousands  of  horse- 
power in  the  service  of  millions  of  people,  and  the  lines  for  some 
of  this  work  are  supported  on  reinforced  concrete  towers.  Electrical 
supply  in  Buffalo,  N.  Y.,  to  the  amount  of  30  000  h.p.,  depends  entirely 
on  the  circuits  from  Niagara  Falls  which  operate  at  22  000  volts  and, 
at  Tonawanda,  N.  Y..  are  supported  on  reinforced  concrete. 

In  the  operation  of  high-voltage  transmissions,  during  the  past, 
some  difficulties  have  been  met,  but  they  have  not  been  so  serious  as  to 
prevent  satisfactory  service.  Nevertheless,  it  is  being  urged  that 
certain  impediments,  met  in  the  operation  of  transmission  systems, 
would  be  much  reduced  by  the  substitution  of  reinforced  concrete  for 
wooden  poles,  and  it  is  even  suggested  that  perhaps  the  first  cost,  and 
probably  the  last  cost,  of  a  transmission  line  of  this  kind  would  be  less 
than  with  wood  for  supports.  The  argument  tor  reinforced  concrete 
in  the  matter  of  costs  is  that,  while  a  tower  requires  a  larger  invest- 
ment than  a  wooden  pole,  yet  the  smaller  number  of  towers  may  reduce 
the  entire  outlay  to  about  the  same  as  for  wood.  More  than  this,  it 
is  said  that  the  lower  depreciation  and  maintenance  charges  on  rein- 


*  Described  by  the  writer  in  his  paper  on  "•  Reinforced  Concrete  Towers,"  Proceedings. 
Am.  Soc.  C.  E.,  Vol.  XXX III,  p.  572. 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  189 

forced  concrete  supports  will  make  their  final   cost  less  than  that  of  Mr.  Krellwitz, 
wooden  poles. 

One  advantage  of  reinforced  concrete  over  wood  is  that  it  will 
not  burn,  and  is  probably  not  siibject  to  destruction  by  lightning. 
The  fact  that  reinforced  concrete  will  not  burn  may  make  it  desirable 
in  places  where  a  long  line  passes  over  a  territory  covered  with  brush 
or  timber.  In  tropical  countries  where  insects  rapidly  destroy  wood 
the  use  of  reinforced  concrete,  even  at  a  much  greater  cost,  might  be 
highly  desirable. 

Guy  B.  Waite,  M.  Am.  Soc.  C.  E.   (by  letter). — Eeinforced  con- Mr.  Waite. 
Crete  has  its  uses,  and,  up  to  the  present,  there  are  few  things  to  which 
it  has  not  been  found  to  apply. 

Public  opinion  has  changed  within  a  very  few  years  from  serious 
doubt  about  concrete  being  good  for  anything  to  that  now  held,  that 
it  is  good  for  everything. 

Friends  of  concrete  can  do  much  damage  to  the  cause  by  insisting 
on  pointing  out  personaj  achievements  where  actually  failures  should 
have  been  recorded. 

It  is  not  possible  for  one  man  to  formulate  a  statement  as  to  the 
universal  adaptability  of  concrete  for  a  given  purpose,  in  all  localities 
from  jSTew  York  to  California,  without  a  knowledge  of  all  the  condi- 
tions in  each  locality.  The  popular  idea  seems  to  be  in  most  places 
that  concrete  should  be  used  for  buildings  because  it  is  so  much 
cheaper  than  wood,  and  that  in  concrete  construction  the  cost  of  almost 
anything  is  very  trifling.  This  view  has  recently  been  strengthened  by 
one  of  our  most  distinguished  and  respected  prophets,  who  promises 
to  see  that  a  two-family  house,  if  it  is  desired,  is  turned  out  complete 
in  a  few  hours.  It  is  to  be  regretted  that  the  necessary  details  to  enable 
others  to  benefit  by  his  discovery  are  not  disclosed. 

Concrete  has  its  pros  and  cons  which  could  be  stretched  in  long 
columns,  thus,  for  example : 

Against  concrete:  For  concrete: 

Xot  good  in  tension;  Good  in  compression; 

Requires  forms;  Good  for  limited  amount  of  shear; 

Eequires  time  to  set;  Strength  improves  with  age; 

Difficult  to  tear  down — or  to  fall  Economical  where  forms  are  sim- 

down ;  pie ; 

Etc.  Is  monolithic; 

Etc. 

Stone  concrete,  mixed  in  the  proportion  of  1:2:4,  can  be  laid 
down  in  almost  any  part  of  a  fair-sized  building,  with  profit,  at  30 
cents  per  cu.  ft.,  not  including  forms. 

An  average  steel  colvimn,  for  a  corresponding  building,  could  be 
erected,  at  a  profit,  for  $90  per  ton. 


190  DISCUSSION  ON  THE  USE  OF  REINFOKCED  CONCRETE     [Papers. 

Mr.  Waite.        Average  steel  floor  beams  and  girders,  of  standard  sections,  will 
cost  $60  per  ton. 

Beginning  with  the  supporting  columns  of  a  building,  a  properly 
reinforced  concrete  column  (conservatively  estimated)  will  carry  an 
average  of  750  lb.  per  sq.  in.  On  the  other  hand,  suppose  the 
corresponding  steel  column  to  be  strengthened  so  that  it  carries  an 
average  of  16  000  lb.  per  sq.  in.  Then  the  required  amount  of  ma- 
terials in  the  two  cases  will  be  as  750  to  16  000,  or  about  as  1  to  21. 

The  costs  of  corresponding  sections  of  the  two  materials,  on  the 
foregoing  assumption,  will  be  30  cents  and  $21.96,  or  as  1  to  73.  There- 
fore the  relative  costs  of  the  sections  of  each  material  to  carry  any 
unit  loading  will  be  as  21  to  73,  or  about  1  to  34  in  favor  of  the  con- 
crete column. 

From  here  on,  practical  experiences  will  become  useful  to  decide 
whether  the  percentage  of  1  to  3J  in  favor  of  concrete  is  the  ultimate 
ratio  of  cost,  when  everything  is  considered. 

Even  engineers  prejudiced  in  favor  of  steel  will  perhaps  concede 
that  for  this  steel  column  about  12  to  15%  will  have  to  be  added  to 
the  carrying  shaft  for  fittings,  etc.  (and  in  the  case  of  latticed  columns 
much  more  than  this),  which  added  amount  of  steel  will  be  sufficient 
to  reinforce  the  concrete  column — according  to  the  accepted  theory  of 
hooping.  Further,  if  the  steel  column  is  to  be  protected  from  rust  as 
well  as  fire,  the  forms  and  the  concrete  material  for  such  fire -proofing 
will  be  substantially  the  same  in  each  case. 

Without  taking  time  to  go  further  into  details,  it  would  appear 
that  concrete  properly  used  in  the  form  of  columns  would  certainly 
have  the  better  of  the  argument,  when  comparing  costs. 

The  speed  of  erection  sometimes  becomes  important,  and,  where 
the  reinforcement  to  the  concrete  column  is  made  in  the  form  of  an 
independent  carrier,  construction  can  proceed  approximately  as  rapidly 
as  in  all-steel  construction. 

The  next  objection  to  the  concrete  column  is  naturally  the 
increased  size.  This  objection  cannot  be  raised  consistently  except  in 
normal  buildings  more  than  six  stories  high,  and  this  in  the  lower 
stories  only.  If  the  buildings  be  eight  stories  high,  the  size  of  the 
columns  will  only  be  abnormally  large  in  the  two  lower  stories,  etc. 
A  well-constructed  building,  six  stories  high,  should  have  columns  of 
steel  of  not  less  than  a  certain  outward  dimension,  in  order  to  give 
proper  rigidity  to  provide  against  eccentric  loading,  etc.,  and  such 
steel  columns,  when  fire-proof,  will  be  substantially  of  the  size  of  the 
solid  reinforced  concrete  column,  with  an  equivalent  strength  and 
rigidity. 

With  development  along  the  lines  of  improved  reinforcement  for 
the  concrete,  in  reinforced  concrete  columns,  it  is  believed  that  in  the 
future  the  sizes  of  concrete  columns  can  be  reduced  to  meet  all  re- 
quirements. 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  191 

Concrete,   in   connection   with  reinforced   floors,   is  usually   taken  ^^'^-  ^*'^''^®- 
with  a  working  stress  of  500  lb.  extreme  fiber  strain. 

"With  the  usual  T-section  of  floor  construction,  an  average  work- 
ing load  on  the  entire  sectional  area  for  compression  can  be  taken 
safely  at  450  lb.  per  sq.  ft.  Estimating  the  steel  beams  to  take  this 
load  at  the  unit  prices  set  forth  above,  the  comparative  costs  of  con- 
crete and  steel  would  be  about  as  35  to  49,  showing  an  economy  in 
favor  of  concrete,  other  things  being  the  same.  But,  in  this  item  of 
floor  construction,  the  concrete  floors  have  to  be  installed,  even  when 
the  all-steel  construction  is  used,  in  order  to  coat  the  steelwork  and 
protect  it  against  rust  as  well  as  fire.  So  that,  in  reality,  if  the  con- 
crete cost  as  much  or  more  than  the  steel  doing  the  compression  work, 
whatever  is  saved  by  putting  this  concrete  to  work  is  a  clear  gain, 
other  things  being  the  same. 

Other  things  do  not  always  remain  the  same,  however,  and  it  is 
necessary  to  consider  the  form  work  for  the  reinforced  concrete  con- 
struction and  the  forms  for  the  fire-proofing,  used  when  steel  con- 
struction carries  all  the  floor  loads. 

With  the  steel  beams  and  girders  giving  the  working  lines  and 
offering  ample  supports  for  the  wood  forms,  the  modern  system  of 
forms  for  fire-proofing  is  very  materially  less  than  where  much  stronger 
independent  framings  and  supports  must  be  carefully  leveled  and  sup- 
ported for  the  reception  of  the  concrete,  in  reinforced  work. 

Wliere  the  forms  can  be  made  in  the  same  general  manner  as  fire- 
proofing  (as  in  some  improved  systems  of  reinforced  concrete),  the 
discussion  of  the  relative  costs  of  forms  can  be  dropped,  and  one  may 
proceed  to  compare  other  items  in  the  relative  costs  of  concrete  and 
steel  constructions.  Now,  assuming  that  forms  are  the  same,  and  that 
the  concrete  is  used  as  a  fire-proofing  in  each  case,  showing  a  gain  for 
every  bit  of  the  concrete  in  the  reinforced  scheme  (which  is  not  ob- 
tained in  the  fire-proof  scheme),  then,  if  it  is  not  clear  that  there  is 
economy  in  the  reinforced  scheme,  it  is  because  the  concrete  can  be 
made  cheaper  in  the  one  construction  than  in  the  other.  The  floor 
slabs  will  have  the  same  loads  to  carry  when  acting  as  carriers  from 
beam  to  beam:  the  concrete,  to  be  an  effective  protection  to  the  steel 
against  deterioration,  must  be  rich,  so  that,  if  the  ultimate  objects 
are  to  be  accomplished,  the  concrete  should  be  substantially  the  same 
in  either  case. 

Without  making  the  inquiry  more  monotonous,  it  would  appear 
that,  in  floors,  concrete  reinforced  construction  shows  an  economy  in 
proportion  to  the  amount  of  steel  it  is  able  to  replace.  So  that,  where 
economy  alone  is  the  object,  a  good  steel  job  is  necessary.  In  light 
constructions  (such  as  dwellings  and  hotels),  where  but  little  steel  is 
necessary,  one  cannot  save  as  much  by  using  concrete  as  where  the 
steel  is  heavier;  and  the  saving  continues  to  increase  with  the  amount 
of  steel  to  be  saved. 


193  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE      [P^ipf'-s^ 

Ml-.  Waite.  In  the  foregoing  comparison  of  relative  costs  in  column  and  floor 
constructions  the  form  work  is  similar  whether  reinforced  concrete  or 
steel  and  fire-proofing  be  used.  In  monolithic  wall  and  partition  con- 
struction the  comparison  is  disadvantageous  when  it  is  considered 
that  brick  walls  and  partitions  are  laid  rapidly  and  without  the  incon- 
veniences of  forms,  and  that  double  forms  are  necessary  for  concrete. 
Further,  it  is  very  much  more  difficult  to  place  the  forms  for  straight 
walls  or  partitions  than  for  either  columns  or  floors.  The  wall  forms 
are  not  easily  held  plumb,  or  in  straight  lines.  The  removal  of  the 
forms  for  walls  is  also  much  more  difficult  than  for  either  columns  or 
floors.  The  cost  of  common  brick  and  mortar  amounts  to  about  18 
cents  per  cu.  ft.,  and  the  cost  of  the  materials  composing  concrete  is 
just  about  the  same.  So  that  the_  cost  of  laying  the  brickwork,  for 
walls  of  the  same  thickness,  must  be  balanced  by  the  cost  of  the  double 
forms  and  placing  the  concrete. 

It  is  not  intended  to  burden  the  reader  with  descriptions  of  the 
difficulties  of  constructing  form  work  for  vertical  structures;  but,  to 
anyone  having  much  experience,  it  must  be  evident  that  such  diffi- 
culties must  be  met.  Economy  in  wall  work  must  be  looked  for  only 
in  heavy  work,  where  the  quantity  of  material  placed  for  any  given 
form  is  sufficient  to  pay  for  it,  without  materially  affecting  the  cost 
of  the  concrete. 

When  no  finish  is  looked  for  on  the  concrete  work,  rough  forms 
may  be  placed  for  from  4  to  5  cents  per  sq.  ft.  on  each  side  of  the  wall; 
but,  for  good  form  work,  the  cost  will  run  from  7  to  10  cents  per  sq. 
ft.  on  each  side. 

Concrete  walls  will  be  erected.  They  are  an  improved  construction, 
and  can  be  handled  conveniently  in  connection  with  other  concrete 
work  in  a  building.  The  object  of  writing  what  seems  to  the  writer 
to  be  the  truth  about  their  construction  is  that  economy  in  their  con- 
struction should  be  looked  for  along  other  lines  than  making  double 
forms  for  the  reception  of  the  concrete.  It  is  believed  that  there  will 
soon  be  other  means  of  erecting  concrete  walls  and  partitions,  in  which 
concrete  can  more  than  compete  with  the  rapid  and  economical  brick 
wall. 
Mr.  siocum.  C.  L.  Sloclim,  Assoc.  M.  Am.  Soc.  C.  E.  (by  letter).— The  science 
and  use  of  reinforced  concrete  in  the  United  States  appears  to  be  in 
its  earlier  stages,  as  compared  with  a  longer  and  more  thorouph  ac- 
quaintance and  varied  use  in  Europe.  Only  recently  its  wide  application 
in  America  has  been  appreciated  in  the  manifold  kinds  of  construction 
which  are  now  seen  almost  everywhere.  Generally  speaking,  theory 
and  practice  do  not  seem  to  be  as  closely  allied  in  America  as  abrond. 
American  engineers  have  not  learned,  as  well  as  European  engineers, 
that  knowledge  of  the  constituent  materials  and  thoroughness  in  de- 
tails of  construction  are  more  important  than  records  in  speed  of  erec- 


Tapers.]     DISCUSSION  ON  THE  USE  OP  REINFORCED  CONCKETE  VJo 

tion.  Like  everything  new,  mneh  opposition,  in  the  nature  of  in-  Mr-  siocum. 
credulity,  has  to  be  overcome.  For  its  age,  reinforced  concrete  is 
fairly  well  understood,  and  it  may  be  said  that  its  newness  is  its 
greatest  fault.  The  change  in  the  field  of  design  caused  by  the 
knowledge  of  the  properties  and  capabilities  of  the  combination  of 
concrete  and  steel  is  now  general,  and  is  somewhat  in  the  nature  of  a 
revolution  in  construction.  There  is  hardly  a  department  or  particu- 
lar sphere  of  construction  which  has  not  been  changed  by  it.  Homely 
and  incongruous  constructions  in  wood,  steel,  and  stone,  and  other 
types  of  construction  too  highly  commercialized,  may  now,  at  reason- 
able, cost,  give  place  to  permanent  structures,  which  are  pleasing  to  the 
eye  and  are  harmonious  additions  to  the  locality  or  latidscape.  Many 
types  of  construction  in  vogue  or  considered  as  good  standard  practice 
two  or  three  years  ago  are  now  appropriately  known  or  should  be 
known  as  a  part  of  the  history  of  construction. 

If  reinforced  concrete  can  be  accorded  the  same  conscientious  treat- 
ment and  scrutiny  as  steel  receives,  there  need  be  no  hesitation  about 
making  the  change  to  more  permanent  and  artistic  structures,  which, 
if  honestly  built,  will  cause  no  concern  or  attention  after  they  are  put 
in  place.  The  mature  design  and  construction  of  steelwork  to-day  is 
accomplished  by  experts  in  that  line,  and  these  are  necessary  accom- 
paniments of  its  age  and  maturity.  The  use  of  reinforced  concrete 
needs  more  rigid  inspection  in  construction,  for  it  is  idle  to  apply  care- 
fully intricate  formulas  to  designs  which  when  constructed  suffer  for 
want  of  expert  superintendence  and  experienced  labor. 

In  the  realm  of  bridge  construction,  where  ample  depth  is  available, 
there  is  not  much  doubt  as  to  its  economy.  This  still  holds  true  for 
spans  with,  comparatively  shallow  depth,  and  with  light  loads,  in  the 
nature  of  moving  concentrations.  For  crossings  with  little  depth  of 
structure  available,  with  heavy  moving  concentrations,  its  sphere  of 
usefvilness  is  at  present  advisedly  confined  to  short  spans.  However, 
even  floor  spans,  up  to  and  from  30  to  40  ft.,  under  heavy  concentra- 
tions, with  less  than  the  ordinary  depth,  can  well  be  investigated.  Fab- 
ricated units,  of  simple  shapes,  as  reinforcement,  with  little  or  no  shop- 
work,  will  afford  ample  stiffness.  Theoretical  analysis,  however,  must 
show  that  the  unit  stresses  in  the  concrete  and  steel  are  well  within 
the  fatigue  limits.  Continuous  framework  or  an  interdependent  sys- 
tem of  units,  easily  put  together,  as  reinforcements,  but  rigid  in  itself 
when  complete,  would  seem  to  afford  as  much  stiffness  as  steel  beams 
bedded  in  concrete,  which  are  generally  calculated  as  carrying  all  the 
loads  independently.  In  true  reinforced  work  the  homogeneous  com- 
bination of  the  concrete  and  steel  is  the  supporting  resistance.  The 
full  use  of  the  two  materials  to  carry  the  loads  must  be  more  economical 
than  the  use  of  the  one  which  has  the  concrete  merely  as  a  protection. 
The  writer  doubts  the  eennniuy  of  hybrid  construction. 


194  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  siocum.  With  old  OP  nmcli--used  material  the  internal  or  molecular  structure 
and  properties  of  which  have  been  changed,  or  are  in  doubt,  half  the 
usual  unit  stress  allowed  for  new  reinforcement,  or  doubling  the  usual 
economical  percentage  of  reinforcement,  would  seem  to  be  safe  and 
advisable.  The  use  of  old  material,  of  cumbersome,  as  well  as  dubious 
section,  of,  say,  4  or  5  sq.  in.  net  section,  such  as  old  rails,  is  inad- 
visable for  floor  bridges  in  total  length  greater  than  the  commercial 
rail  lengths;  because  attempts  to  develop  such  sections  in  tension  are 
too  expensive,  and  are  somewhat  abortive. 

In  a  series  of  short,  independent,  self-supporting  arches  of  rein- 
forced concrete,  which  are  very  flat,  and  are  practically  carried  on  con- 
tinuous columns,  the  writer  has  used  the  cantilever  method  in  finding 
the  stresses  in  the  constituent  materials,  and  has  proportioned  the 
steel  accordingly;  in  other  words,  he  has  considered  the  middle  third 
of  each  span  as  carried  by  the  end  thirds.  These  arches  were  calcu- 
lated for  the  heaviest  moving  concentrations  for  highways.  In  beam 
and  slab  bridges,  carrying  heavy  trolley  concentrations,  where  the  de- 
sign is  somewhat  hampered  for  depth,  higher  percentages  of  steel  and 
double  reinforcement  may  have  economical  advantage. 

In  current  American  practice,  more  time  can  be  allowed  to  good 
advantage  for  this  construction  to  attain  mature  strength  rather  than 
use  a  green  structure  prematurely  and  perhaps  lessen  the  efficiency  of 
the  bond.  Collections  of  materials  of  construction  or  equipment, 
sometimes  inadvertently  placed  on  new  work,  give  concentrations  for 
which  the  design  is  not  calculated,  and,  if  the  work  is  not  of  sufficient 
age,  much  damage  may  be  done,  and  may  not  be  evident  until  some 
time  after.  Such  consequent  weakness  may  be  brought  out  by  fatigue, 
which,  under  normal  conditions,  could  not  be  explained.  From  ob- 
servation, competent,  well-paid  superintendence  and  experienced  work- 
men of  the  best  class  give  the  strongest  structure  and  the  one  that 
fulfills  all  the  conditions  of  economy. 

As  compared  with  the  usual  heavy  masonry  arches  of  gravity  sec- 
tion, the  comparatively  light  reinforced  arches  give  more  and  greater 
vibrations  under  moving  loads,  principally  on  account  of  much  less 
bulk  weight  of  structure.  Can  reinforced  concrete  work  vibrate  with 
the  same  impunity  as  steelwork?  The  writer  thinks  it  can,  if  the 
working  stresses  are  not  too  high,  but  are  well  within  the  fatigue  limits. 
Much  interesting  and  instructive  information  could  be  obtained  by 
measuring  the  number  of  vibrations  and  their  amplitude  on  bridges  of 
different  types  under  different  kinds  and  speeds  of  rolling  loads. 
Under  any  conditions,  crossings  of  shallow  floor  construction  can  well 
be  tested  for  unusual  loads,  and  consequent  deflection,  if  any. 

Other  properties  and  characteristics  being  satisfactory,  a  greater 
proportion  of  finely-ground  cement,  with  a  graded  aggregate  will,  with 
safety,  give  reinforced  concrete  design  and  construction  its  bold  quality. 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  195 

which  distinguishes  it.  Too  little  attention  is  paid  to  the  compact-  M'-  siocnm. 
ness  or  density  of  the  mixture.  The  result  of  a  few  simple  and  inex- 
pensive experiments  in  the  measurement  of  voids,  taking  a  compara- 
tively short  time  to  perform,  will  give  a  cheaper  and  stronger  concrete. 
In  reinforced  work,  such  preliminary  investigations  are  productive  of 
economy. 


196  MEMOIR  OF   CHARLES  PAINE  [Memoirs. 

MEMOIRS  OF  DECEASED  MEMBERS. 

Note.— Memoirs  will  be  reproduced  in  the  volumes  of  Transactions.  Any  information 
which  will  amplify  the  records  as  here  piiuted.  or  correct  any  errors,  should  be  forwarded 
to  the  Secietary  prior  to  the  final  publication. 


CHARLES  PAINE,  PAST-PRESIDENT,  AM.  SOC.  C.  E. 


Died  July  4th,  1906. 


Charles  Paine  was  born  in  Haverhill,  New  Hampshire,  on  April  25th, ' 
1830,  and  died  at  Tenafly,  New  Jersey,  on  July  4th,  1906.  Almost  his 
entire  working  life,  which  begdn  when  he  was  14  years  old,  was  spent 
in  railroad  service,  and  the  building  of  his  professional  position  was 
in  that  period  when  the  railroad  art  was  still  primitive  and  when  the 
opportunities  for  education  as  a  Civil  Engineer,  in  the  United  States, 
were  confined  almost  to  actual  work  in  the  office  and  in  the  field. 

Mr.  Paine  was  descended  from  Stephen  Paine,  who  came  to  the 
United  States  from  England  in  1638,  and  his  family  remained  in  New 
England  for  eight  generations.  He  belonged  to  a  fine,  substantial 
stock,  and  some  of  his  ancestors  were  distinguished. 

Mr.  Paine's  school  education  was  quite  limited.  In  1839,  he  en- 
tered the  College  of  the  Order  of  St.  Sulpice,  in  Montreal,  where  he 
remained  for  two  years  and  where  he  acquired  that  part  of  his  school 
education  which  always  seemed  to  him  the  most  valuable,  viz.,  a  good 
grounding  in  French  and  Latin  and  in  the  elements  of  what  old- 
fashioned  people  call  "polite  education."  He  spent  a  year  at  an 
academy  in  Meriden,  New  Hampshire,  and  two  years  in  New  York  at 
a  school  of  which  Mr,  Charles  Coudert  was  principal.  This  Mr. 
Coudert  was  the  father  of  the  late  Charles  and  Frederic  Coudert, 
eminent  lavpyers  of  New  York  and  Paris.  From  an  uncle,  William  T. 
Porter,  editor  and  proprietor  of  the  New  York  Spirit  of  the  Times, 
young  Paine,  during  these  two  years,  was  permitted  certain  oppor- 
tunities to  see  life  and  people,  and  he  always  attached  a  good  deal  of 
value  (and  not  without  reason,  perhaps)  to  the  hours  which  he  passed 
in  the  offices  of  that  newspaper  in  the  company  of  the  wits,  men  about 
town,  and  famous  actors  and  actresses.  He  says  that  in  that  brief 
period  he  "drank  in  a  love  of  fine  things  in  conduct,  in  art,  in  litera- 
ture, and  in  manners  which  has  continued  a  joy  to  me  throughout  my 
life." 

In  the  spring  of  1844,  Paine's  uncle,  Governor  Charles  Paine,  of 
Vermont,  took  the  lad  into  the  counting-room  of  his  broadcloth 
factory,  where  he  remained  until  August,  1845.  Then  he  entered  the 
service  of  the  Vermont  Central  Railroad,  the  surveys  for  which  had 
just  been  commenced.     Of  this  enterprise.  Governor  Paine  was  Presi- 

*  Memoir  prepared  by  Edward  P.  North  and  H.  G.  Prout,  Members,  Am.  Soc.  C.  E. 


Memoirs.]  .    MEMOIR  OF  CHARLES  PAINE  1^? 

dent.  Here  he  began  as  rodman  in  the  corps  of  Charles  Brown,  one 
of  the  old  engineers  of  the  period,  who  enjoyed  a  fine  reputation. 
Other  young  men,  afterward  distinguished,  who  were  associated  with 
Paine  in  this  work  at  this  time,  were  the  late  S.  M.  Felton,  M.  Am. 
Soc.  C.  E.,  afterward  President  of  the  Philadelphia,  Wilmington  and 
Baltimore  Eailroad  and  of  the  Pennsylvania  Steel  Company;  Mr. 
Charles  Collins,  afterward  Chief  Engineer  of  the  Lake  Shore  and 
Michigan  Southern;  Mr.  Carpenter,  afterward  United  States  Senator 
from  Wisconsin;  and  Dr.  Edward  H.  Williams,  afterward  General 
Superintendent  of  the  Pennsylvania  Division  of  the  Pennsylvania 
Railroad,  and  a  member  of  the  firm  of  Burnham,  Parry,  Williams  and 
Company,  owners  of  the  Baldwin  Locomotive  Works.  Mr.  Paine  en- 
joyed the  close  friendship  of  all  these  men  until  their  lives  ended,  and 
it  is  related  that  at  the  time  of  the  Chicago  fire,  in  1871,  Dr.  Williams 
and  his  wife,  knowing  that  the  Paines  lived  in  Chicago,  immediately 
shipped  by  express  from  Philadelphia  a  complete  outfit  of  clothes  for 
each  member  of  the  Paine  family,  without  stopping  to  ask  if  the  clothes 
were  needed. 

In  the  autumn  of  1847,  the  Vermont  Central  was  completed  into 
Northfield,  and,  for  a  very  short  time,  young  Paine  got  a  chance  to 
fire  a  locomotive,  which  he  always  regarded  as  one  of  the  most  valu- 
able experiences  of  his  life.  The  following  winter  he  spent  in  the 
drafting  rooms  of  Brown  and  Hastings,  Civil  Engineers,  in  Boston, 
and  of  Hinkley  and  Drury's  Locomotive  Works,  where  he  made  quite 
complete  drawings  of  all  the  parts  of  a  locomotive. 

Li  1848,  Mr.  Paine  took  charge  of  a  division  of  the  Vermont  and 
Canada  Railroad,  under  Henry  R.  Campbell,  Chief  Engineer.  This 
road  was  completed  in  1850.  Mr.  Paine  then  went  to  Montreal  and 
took  charge  of  the  contracts  of  H.  R.  Campbell  for  building  a  rail- 
road from  Rouses  Point  to  St.  Johns,  and  for  building  a  branch  line 
from  St.  Lambert  to  intersect  with  the  line  of  railway  between  St, 
Johns  and  La  Prairie.  At  this  time  he  also  had  charge  of  the  build- 
ing of  docks  at  Moffatt's  Island,  opposite  Montreal. 

It  will  be  seen  that  the  young  man  had  considerable  responsibili- 
ties before  he  was  of  age,  and  he  appears  to  have  been  in  no  way  re- 
luctant to  assume  still  other  responsibilities,  for  on  May  13th,  1851, 
less  than  a  month  after  reaching  his  majority,  he  was  married  to 
Olivia  Blodgett  Hebard,  of  Chelsea,  Vermont.  His  wife  belonged  also 
to  one  of  the  most  solid  New  England  families.  She  was  a  woman 
of  great  cultivation  of  mind  and  of  strong  and  beautiful  character, 
and  they  lived  together  in  the  greatest  happiness  until  Mrs.  Paine's 
.death  in  the  summer  of  1897.  They  had  six  children,  and  four  sons 
now  survive. 

In  1855,  Mr.  Paine  moved  to  Wisconsin,  where  he  became  Chief 
Engineer  of  the  Beaver  Dam  and  Baraboo  Railroad,  and  of  the  Fox 


198  MEMOIR  OF  CHARLES  PAINE  [Memoirs. 

Eiver  Valley  Eailroad,  neither  of  which  enterprises  got  beyond  the 
stage  of  grading  the  roadbed,  because  of  the  great  panic  of  1857.  In 
August,  1858,  Mr.  Paine  became  Superintendent  of  the  Western  Di- 
vision of  the  Michigan  Southern  and  Northern  Indiana  Eailroad, 
which  road  was  at  that  time  five  months  behind  in  its  pay-roll  and 
physically  pretty  nearly  a  wreck.  The  local  nickname  for  the  road 
was  the  "Miserably  Slow  and  Nearly  Insolvent  Eailroad."  These  con- 
ditions, however,  were  not  peculiar  to  that  railroad  in  the  year  1858. 
Mr.  Paine's  connection  with  this  railroad  and  its  lineal  successors 
continued  for  twenty-three  years.  In  January,  1864,  he  was  made 
Chief  Engineer  of  the  road,  and  on  March  1st,  1872,  he  became  Gen- 
eral Superintendent  of  the  Lake  Shore  and  Michigan  Southern"  Eailway. 
While  in  charge  of  this  road  he  made  such  improvements  and  econo- 
mies that  by  1876  he  had  demonstrated  his  ability  to  carry  freight  for 
4  mills  per  ton-mile,  and  from  this,  at  the  time,  sriiall  sum,  pay  all 
the  costs  except  for  improvements,  dividends,  and  interest  on  the 
bonded  debt. 

He  remained  Superintendent  of  the  Lake  Shore  and  Michigan 
Southern  until  he  was  appointed  General  Manager  of  the  New  York, 
West  Shore  and  Buffalo,  in  August,  1881.  He  organized  and  carried 
through  the  building  of  this  road,  and  upon  its  bankruptcy  he  found 
himself  with  health  impaired  and  with  the  savings  of  his  lifetime 
gone,  for  he  himself  had  invested  in  the  securities  of  the  enterprise  in 
which  he  believed  enthusiastically. 

In  order  to  get  himself  in  condition  to  rebuild  his  fortunes,  he 
adopted  the  novel  and  bold  scheme  of  traveling  in  Europe  for  a  year 
on  borrowed  money.  The  remedy  was  characteristic  and  highly  suc- 
cessful, and  until  the  day  of  his  death  he  never  suffered  another  illness. 

He  served  for  a  short  time  as  the  General  Superintendent  of  the 
New  York,  Pennsylvania  and  Ohio  Eailroad,  and  for  a  few  months 
as  Second  Vice-President  of  the  Erie,  and  then  he  went  to  Pittsburg 
to  help  Mr.  Westinghouse  in  developing  the  natural  gas  industry 
through  the  Philadelphia  Company.  There  he  remained  until  Decem- 
ber, 1890,  he  having  had  active  executive  charge  of  the  company. 

He  returned  to  New  York  at  the  end  of  1890,  and  opened  an  office 
as  Consulting  Engineer,  which  office  he  maintained  until  1899;  part  of 
which  time,  however,  he  was  General  Manager  of  the  Union  Steam- 
boat Line,  a  subsidiary  Erie  company,  and  he  occupied  an  important 
and  confidential  position  in  the  administrative  organization  of  the  Erie. 

From  1899  until  a  year  before  his  death,  Mr.  Paine  was  General 
Manager  of  the  Panama  Eailroad  Company,  and  for  a  time  he  was 
also  Vice-President  and  a  Director  of  that  Company.  This  service 
ended  with  the  purchase  of  the  Panama  Eailroad  by  the  United  States 
Government  and  the  transfer  of  its  management  to  the  existing  Canal 
Commission. 


Memoirs.]  MEMOIR  OF   CHARLES  PAINE  199 

Mr.  Paine  was  elected  a  Member  of  the  American  Society  of 
Civil  Engineers  on  its  reorganization  in  December,  1867,  and  was  the 
second  man  to  join  the  Society :  numbering  17  on  the  list  of  members 
as  it  stood  for  the  first  year.  He  contributed  to  the  Transactions 
Paper  XX :  "History  of  the  Iron  Rails  on  the  Michigan  Southern  and 
Northern  Indiana  Railway,"  and  was  President  of  the  Society  during 
1883. 

At  the  time  of  his  death,  Mr.  Paine  was  a  Member  of  the  Century 
Club,  in  New  York,  an  Honorary  Member  of  the  Western  Society 
of  Engineers,  and  of  the  Engineers'  Club  of  Cleveland,  and  a  number 
of  other  scientific  and  philosophical  bodies. 

Mr.  Paine's  personality  was  so  extraordinary,  and  meant  so  much 
to  those  among  whom  he  lived,  that  special  mention  should  be  made 
of  it.  His  manner  was  commanding,  but  singularly  gracious.  He  had 
a  dignified  and  impressive  presence.  He  was  of  generous  and  en- 
thusiastic temperament.  He  had  a  broad  sympathy,  wide  reading,  and 
a  discriminating  taste  in  literature  and  art;  but,  beyond  this,  there 
ib  much  more  to  be  said.  In  every  generation  there  are  a  few  men 
who  impress  their  fellow  men  by  beauty  and  nobility  of  character, 
quite  apart  from  those  qualities  which  we  may  think  of  as  purely 
intellectual.  They  have  a  distinction  which  wealth  or  power  or  achieve- 
ment cannot  bestow.  In  the  deepest  recesses  of  our  minds  we  recognize 
these  men  as  being  the  real  nobility — the  flower  of  humanity.  Mr. 
Paine  belonged  to  the  small  group  of  men  distinguished  by  character. 
He  had  intellectual  superiority,  and  he  was  a  man  of  honorable  achieve- 
ment; but  we,  who  knew  him  well,  think  of  him  first  and  respect  him 
most  for  the  subtle  qualities  of  gentle  manliness.  His  temper  was 
naturally  quick,  and  he  had  great  personal  dignity;  but  his  courtesy 
was  unfailing  and  his  modesty  was  sincere.  He  was  chivalric  in 
thought  and  conduct.  Honor,  truth,  and  duty  were  in  the  roots  of  his 
nature — inherited,  bred  in  the  bone.  These  were  his  shining  charac- 
teristics, by  virtue  of  which  his  life  was  lived  in  a  high  and  serene 
atmosphere,  and  in  that  atmosphere  dwelt  with  him  a  wife,  his  equal 
in  every  way. 


O^QCo 


AMERICAN   SOCIETY 


r 


OF 


CIVIL  ENGINEERS 


March,    1908. 
PROCEEDINGS  -  VOl!  XXXIV—No.  3 


N.  L     N^/,L 

E:nginee:ring  Society 

William   p.   Morse 


Published  at  the  House  of  the  Society,  220  West  FItty-scventh  Street,  New   York 
the  Fourth  Wednesday  of  each  Month,  except  June  and  July. 

Copyrighted  1908,  by  the  American  Society  of  Civil  Engineers. 

Entered  as  Second-Class  Matter  at  the  New  York  City  Post  Office,  December  15th,  1896. 

Subscription,  S6  per  annum. 


Vol.    XXXIV.  MARCH,  1908.  No.  3. 


AMEEICAN  SOCIETY  OF  CIVIL  ENGINEEES. 

INSTITUTED    185  3, 


PAPERS  AND   DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


CONTENTS. 

Papers :  page. 

Erection  of  the  Bellows  Falls  Aich  Bridge. 

By  L.  D.  Rights.  Assoc.  M.  Am.  Soc.  C.  E 303 

Recent  Developments  in  Pneumatic  Foundations  for  Buildings. 

By  D.  A.  UsiNA,  Assoc.  Am.  Soc.  C.  E 313 

Substructure  of  Piscataquis  Bridge,  and  Analysis  of  Concrete  Work. 

By  G.  A.  Hersey,  Jun.  Am.  Soc.  C.  E .' 323 

Discussions  : 

Overhead  Construction  for  High-Tension  Electric  Traction  or  Transmission. 

By  Messrs.  Joseph  Mayer,  W.  K.  Archbold,  Charles  Rufus  Harte,  Farley 

Osgood,  and  W.  S.  Murray 239 

A  New  Suspension  for  the  Contact  Wires  of   Electinc  Railways  Using  Sliding 
Bows. 

By  Messrs.  R.  D.  Coombs  and  Charles  Rufus  Harte 354 

Safe  Stresses  in  Steel  Columns. 

By  Messrs.  Henry  B.  Seaman,  Luzerne  S.  Cowles,  Charles  M.  Emmons, 
Henry  S.  Prichard,  Horace  E.  H(>rton,  F.  P.  Shearwood,  L.  D.  Rights,  and 
A.  W.  Carpenter 357 

Effect  of  Earthquake  Shock  on  High  Buildings. 

By  Messrs.  Uuy  B.  Waite,  and  E.  G.  Walker 393 

The  Use  of  Reinforced  Concrete  in  Engineering  Structures:  An  Informal  Discus- 
sion. 
By  Messrs.  M.  S.  Falk,  Rudolph  P.  Miller,  Eugene  W.  Stern,  and  H.  C. 
Turner 397 

IVlemoirs: 

Calvin  Easton  Brodhead,  M.  Am.  Soc.  C.  E 308 

Georoe  Thomas  Nelles.  M.  Am.  Soc.  C.  E 309 

Herbert  Franklin  Northrup,  M.  Am.  Soc.  C.  E 311 

William  Roberts,  Assoc.  Am.  Soc.  C.  E 313 

PLATES. 

Plate  XXVIII.        Strain  Sheet,  Bellows  Falls  Arch  Bridge 205 

Plate  XXIX  Erection  of  Bellows  Falls  Arch  Bridge 207 

Plate  XXX.  Method  of  Erecting  Bellows  Falls  Arch  Bridge 309 

Plate  XXXr.  Views  of  Bellows  Falls  Arch  Bridge 311 

Plate  XXXII.         South  Abutment,  Piscataquis   Bridge,  Showing  Method  of  Placing 

Concrete,  and  Coffer- Dam  of  Pier  3.  Showing  Cableway 325 

Plate  XXXIII.       Piscataquis  Bridge,  Looking  South,  Shovving  Progress  of  Work,  and 

Genei-al  View  of  Completed  Bridge 237 

Plate  XXXIV.       Single  Catenary  Construction,  Syracuse,  Lake  Shore  and  Northern 

Railroad '. 239 

Plate  XXXV.  Insulators  Before  and  After  Testing 7 241 

Plate  XXXVI.        Sleet  Accretion  on  Twig  and  on  Wires,  Winsted,  Conn 343 

Plate  XXXVII.      Wire  Gridiron   Under    33  000-Volt  Transmission    Line,    and  Wire 

Cradle  Over  II  000-Volt  Transmission  I,ine 245 

Plate  XXXVIII.    Wood  Bar  Cradle  Under  6  600- Volt  Lighting  Circuit,  and  Catenary 

Crossing,  with  Single  Cut-Off  Arms 249 

Plate  XXXIX.       Improvised    "Atwood's    Machine"     for  Testing    the  Stiffness   of 

Trolley  Wire 255 

Plate  XL.  Reinforced  Concrete  Structure  for  Ice  Storage,  and  Scaffolding  to 

Support  Reinforcing  Rods 399 

Plate  XLI.  Reinforcement  of    Walls  and  Columns,  and  Reinforced  Concrete 

Building 801 

Plate  XLII.  Reinforced  Ccncrete  Construction 303 


Vol.  XXXIV.  MARCH,  1908.  No.  3. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED     1852. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


ERECTION  OF  THE  BELLOWS  FALLS 
ARCH  BRIDGE. 


By  L.  D.  Rights,  Assoc.  M.  Am.  Soc.  C.  E. 
To  BE  Presented  April  1st,  1908. 


The  highway  bridge  across  the  Connecticut  River  at  -Bellows  Falls, 
Vt.,  is  interesting  because,  in  the  United  States,  it  stands  alone  as  an 
example  of  a  through  arch  with  suspended  floor,  and  also  because,  as 
an  arch,  it  is  only  surpassed  in  span  by  the  two  deck  arches  at 
Niagara. 

The  residents  of  North  Walpole,  N.  H.,  depend  largely  on  the 
factories  at  Bellows  Falls,  Vt.,  for  their  employment,  and  on  the  stores 
for  their  trading.  To  reach  the  town,  they  were  compelled  to  use  the 
old  wooden  toll  bridge  at  the  south  end,  or  venture  on  the  Sullivan 
Coimty  (Boston  and  Maine)  Railroad  bridge,  or  patronize  a  rather 
uncertain  rowboat  ferry.  For  years  they  had  urged  a  more  con- 
venient crossing,  and  this  was  naturally  btu;ked  up  by  the  merchants 
and  business  men  on  the  Vermont  side.  The  depth  of  the  river  at  this 
point,  about  25  ft.,  strong  objections  to  piers  above  the  mouth  of  the 
canal,  owing  to  the  vested  rights  of  the  Canal  Company,  and  the  free- 
ing of  the  old  toll  bridge  before  another  could  become  available,  were 
factors  contributing  to  the  delay  of  the  project,  which  resolved  itself, 
largely,  into  a  matter  of  cost. 

Note. — 'These  papers  are  issued  before  the  date  set  for  presentation  and  discussion. 
Correspondence  is  invited  from  those  who  cannot  be  present  at  the  meeting,  and  may  be 
sent  by  mail  to  the  Secretary.  Discussion,  eidier  oral  or  written,  will  be  published 
in  a  subsequent  number  of  PrnceefiingR.  and  when  finally  closed,  the  papers,  with 
discussion  in  full,  will  be  published  in  Transactions. 


Papers.] 


EKECTION  OV  ARCH   BRIDGE 


20J 


Early  in  the  spring  of  1904,  the  agitation  was  again  revived,  and 
interest;ed  citizens  brought  forward  new  and  old  schemes.  One  that 
met  with  considerable  favor  was  to  locate  a  pier  on  the  rock  in  shallow 
water  just  above  the  angle  of  the  dam,  shown  on  the  map,  Fig.  1. 
This  permitted  the  use  of  two  spans,  but  made  it  necessary  that  these 
spans  should  be  at  an  angle  with  each  other  in  order  to  reach  con- 
venient landing  places  on  the  shores. 


BELLO>V 


In  March,  the  two  towns,  at  their  annual  meeting,  voted  appropria- 
tions to  cover  the  cost  of  freeing  the  old  toll  bridge  and  building  a  new 
free  bridge,  and  appointed  a  joint  committee  to  receive  bids  and  enter 
into, a  contract  for  the  work.  The  committee  secured  some  preliminary 
estimates  on  the  various  schemes,  but,  owing  to  the  probability  that 
the  new  bridge  might  be  used  in  the  future  for  electric  cars,  it  did  not 
favor  the  plan  for  two  spans  at  an  angle,  but  preferred  a  single  span. 


204  ERECTION  OF  ARCH  BRIDGE  [Papers. 

As  the  preliminary  estimates  for  a  single-span  bridge  were  unsatis- 
factory, some  members  of  the  committee  visited  Boston  and  appealed 
to  the  President  of  the  Boston  and  Maine  Railroad,  who  was  inter- 
ested to  the  extent  of  freeing  the  railroad  bridge  from  unauthorized 
foot  passengers,  and  who  responded  by  offering  the  services  of  J.  P. 
Snow,  M.  Am.  Soc.  C.  E.,  in  an  advisory  capacity.  Mr.  Snow  was 
naturally  familiar  with  the  general  surroundings,  but  did  not  feel 
that  conditions  warranted  the  preparation  of  an  elaborate  design; 
therefore,  he  drew  up  general  specifications,  and  called  for  bids,  re- 
questing each  bidder  to  submit  his  own  plans.  Among  the  designs 
submitted  were  several  for  truss-  and  suspension  bridges,  but  all  the 
prices  were  greater  than  the  appropriation,  and  the  bids  were  rejected. 
Mr.  Snow  was  satisfied  that  a  bridge  could  be  built  within  the  specified 
sum,  and  recommended  the  employment  of  J.  R.  Worcester,  M.  Am. 
Soc.  C.  E.,  which  suggestion  the  committee  accepted,  Mr.  Worcester 
concluded  to  adopt  an  entirely  different  type  of  bridge,  and  decided 
that  a  three-hinged,  riveted  arch  with  suspended  floor  would  be  the 
most  artistic  and  suitable  structure  for  the  location.  He  drew  plans 
and  specifications,  and,  on  his  recommendation,  separate  bids  were 
asked  for  the  masonry  and  structural  steel.  The  results  of  the  com- 
petition were  satisfactory,  as  several  bids  were  received  which  were 
within  the  appropriation.  On  the  recommendation  of  Mr.  Snow,  the 
contract  for  the  masonry  was  awarded  to  Joseph  Eoss  and  Sons,  of 
Boston,  and  the  superstructure  to  Lewis  F.  Shoemaker  and  Company, 
of  Philadelphia  and  New  York. 

Design. — As  will  be  seen  by  the  general  plan  and  stress  sheet,  Plate 
XXVIII,  the  bridge  is  about  650  ft.  long,  and  consists  of  a  single,  three- 
hinged,  arch  span,  540  ft.  from  center  to  center  of  end  pins,  with 
a  short  truss  span  at  the  west  end,  104  ft.  8  in.  from  center  to  center 
of  bearings.  This  short  span  was  necessary,  in  order  to  carry  the 
street  over  the  Rutland  Railroad.  It  will  be  noted  that  the  roadway 
is  on  a  grade  of  3.33%,  running  downward  from  the  short  span  to  the 
abutment  at  the  east  end.  The  height  of  the  main  arch  is  90  ft.  be- 
tween the  hinge  centers.  The  truss  chords  follow  the  lines  of  two 
parabolas  14  ft.  apart.  In  order  to  secure  simplicity  of  detail,  the 
trusses  do  not  diverge  at  the  bottom,  but  stand  in  parallel  vertical 
planes,  30  ft.  from  center  to  center.  This  provides  for  a  roadway,  20 
ft.  clear,  and  one  sidewalk,  6  ft.  wide,  as  shown  by  the  cross-section. 


PAPERS,   AM-  SOC.  C.   E. 

MARCH,   1908, 

RIGHTS  ON 

ERECTION  OF  ARCH  BRIDGE. 


HALF  TOP  CHORD  PLAN 


♦ 

Foldout 

Here 
♦  ♦ 

♦ 


PLATE  XXIX. 

PAPERS,   AM.  SOC.   C.   E. 

MARCH,   1908. 

RIGHTS  ON 

ERECTION  OF  ARCH  BRIDGE. 


1 

u 

,.^  ^ 

^ 

^          l^i 

1    1 

^ 

m  —  -" 

Fig.  1.— Bellows  Falls  Arch  BBitiGE  in  Progress. 


-^5^^^^ 

^^^^  \if^ 

^J^ ' 

Sii^          y,ie«i 

— -^^j-.'m^lKk 

ei^:aamv^si 

^N 

U^  f^ 

^  1                  / 

I  Vv^ 

/ 

/ 

/ 
/ 

/ 

Fig.  2. -Connecting  Arch  at  tre  Center. 
(Camera  pointed  upward  at  an  angle  of  15  degrees.) 


Pa  per; 


EKKCTION   OF  AHCH   BRIDGE 


2or 


number  and  spacing  of  tlio  pile  supports.  The  details  of  the  center 
bents,  F,  F-^,  and  F...  are  shown  by  Fig.  2,  which  gives  the  size  of  the 
main  posts  and  bracing,  and  indicates  the  splices  and  number  of  bolts. 

.„-    .....  DETAIL  OF 

BENTS  F.  Fj.    AND  F„ 
Bents  F  and  F^  are  as  shown. 
Bent  F^  has  outside  post  only, 
and  no  bracing. 


2-8"i  12"i  31)' 


-44'6" 


Fig.  3. 

It  was  considered  advisable,  on  account  of  the  switching  facilities, 
to  unload  the  material  on  the  Vermont  side  of  the  bridge  and  carry 
those  pieces  required  for  the  east  side  on  a  standard-gauge  service 
track  running  between  the  towers,  along  the  center  line  of  the  bridge. 
A  10-ton  stiff-leg  derrick  was  placed  on  the  west  shore,  to  unload  ma- 
terial from  the  cars  and  transfer  it  to  the  trucks  on  the  service  track. 
Two  30-h.p.  hoisting  engines,  with  two  drums  and  four  spools  each, 
were  located  at  Bent  E  to  raise  the  steel. 


208  ERECTION  OF  ARCH  BRIDGE  [Papers. 

Erection. — The  masonry  plans  provided  for  piles  to  be  driven  in 
the  foundations  for  the  arch,  and,  as  the  masonry  contractor  was  well 
equipped  to  do  the  work,  the  contract  for  furnishing  and  driving  the 
piles  for  the  falsework  was  sublet  to  him.  The  water  has  an  average 
depth  of  about  25  ft.,  and  the  bottom  is  hard  gravel.  Spruce  piles 
were  specified,  in  order  that  they  might  be  sold  to  the  pulp  mills  after 
the  work  was  completed.  They  were  driven  from  8  to  10  ft.  into  the 
bottom,  and  were  cut  off  and  capped  about  3  ft.  above  the  low-water 
line.  The  falsework  towers  were  completed  in  November,  and  the  shoes 
were  set  on  December  6th,  1904, 

Two  gangs  were  started,  one  from  each  end  of  the  arch,  and  the 
rivalry  between  them  helped  not  a  little  in  the  rapid  erection  of  the 
work.  The  severe  weather  of  the  winter  of  1904-05  will  no  doubt  be 
remembered,  but,  even  in  that  latitude,  there  were  breaks  in  the  cold, 
and  on  two  separate  days  considerable  rain  fell.  The  fear  that  the 
ice  might  go  out,  which  would  mean  taking  out  the  falsework  and 
everything  with  it,  was  a  constant  incentive  to  hasten  the  erection  in 
every  way  possible.  It  had  not  been  the  intention  to  do  any  work  on 
the  ice,  but  after  it  had  frozen  to  the  thickness  of  about  2  ft.,  it  was 
found  to  be  very  convenient  when  assembling  the  chords,  which  were 
handled  in  two  sections  of  four  panels  each.  The  ice  also  acted  as  a 
hindrance,  for  when  the  canal  gates  were  closed,  on  Sundays  and 
holidays,  the  river  rose  about  18  in.,  and  it  was  necessary  to  keep  the 
ice  chopped  free  from  the  piles. 

When  the  ice  first  began  to  form  about  the  falsework,  the  structure 
showed  a  tendency  to  move  down  stream.  This  was  carefully  noted, 
account  of  it  being  taken  in  placing  the  steel.  When  the  arch  was 
swung,  some  of  the  bents  were  found  to  have  moved  down  stream  about 
4  in.  As  the  load  was  put  on  the  falsework,  some  of  the  bents  sank 
slightly,  but  this  settlement  was  adjusted  with  wedges  under  the  block- 
ing at  each  point  of  support. 

The  two  end  panels  of  the  lower  chord  and  the  end  panel  of  the 
upper  chord  were  shipped  riveted  together.  These  members,  each 
weighing  about  8  tons,  constituted  the  heaviest  pieces  to  be  handled. 

In  beginning  the  erection,  the  shoes  and  end  panels  of  the  west 
end  were  set  with  the  stiff-leg  derrick  used  for  handling  material  from 
the  siding  to  the  material  trucks.  On  the  east  end  a  gin  pole  was  placed 
to  set  the  shoes  and  end  panels.    Provision  had  been  made  in  the  plans 


PLATE  XXX. 

PAPERS.  AM.  80C.  C.   E. 

MARCH.  1908. 

RIQHT6  ON 

ERECTION  OF  ARCH  BRIDGE. 


BELLOWS  FALLS.  VT 


ROCKINGHAM,  VT, 


iraffi 


♦ 

Foldout 
Here 

♦  ♦ 

♦ 


C.   E. 


IRIDGE. 


I 


i> 


PLATE  XXXI. 

PAPERS,  AM.  SOC.  C.  E. 

MARCH,   1908. 

RIGHTS  ON 

ERECTION  OF  ARCH  BRIDGE. 


F'iG.  1.— The  Bellows  Fails  Arch  Bridge. 


Fig.  2.— End  View, -Bellows  Falls  Arch  Bridge. 


Papeis.]  ERECTIOX   OF  AKCH   BRIDGE  211 

Credit  for  the  work  should  be  given  to  Mr.  J.  H.  Fichthorn,  Chief 
Engineer  for  Lewis  F.  Shoemaker  and  Company,  and  to  Mr.  A.  L. 
Westbrook,  Field  Superintendent.  The  writer  also  wishes  to  ex- 
press his  thanks  to  Messrs.  J.  P.  Snow  and  J.  R.  Worcester  for  the 
information  furnished  for  the  preparation  of  this  paper. 


Vol.  XXXIV.  MARCH,  1908.  No.  3. 


AMEEICAN   SOCIETY  OF  CIVIL  ENGINEEES. 

INSTITUTED     185  2. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


KECENT   DEVELOPMENTS   IN   PNEUMATIC 
FOUNDATIONS  FOR  BUILDINGS. 


By  D.  a.  Usina,  Assoc.  Am.  Soc.  C.  E. 
To  BE  Presented  April  15th,  1908. 


The  purpose  of  this  paper  is  to  review  briefly  the  recent  and  very 
interesting  development  in  foundations  of  the  class  generally  used  for 
the  high  buildings  being  erected  in  the  lower  section  of  New  York  City. 
The  earth  there  overlies  a  stratum  of  rock,  the  depth  of  which  varies 
from  40  to  100  ft.,  and  the  enormous  loads  are  carried  most  securely 
by  concrete  piers  built  with  pneumatic  caissons,  and  resting  directly 
on  the  substratum  of  rock. 

Prior  State  of  the  Art. — Prior  to  the  present  improvements,  the 
conventional  type  of  construction  was  as  ilkistrated  in  Figs.  1  and  2. 
The  working  chamber  was  built  with  sides  and  roof  of  heavy  timber 
or  of  sheet  steel  with  stiifeners  at  suitable  intervals.  The  coffer-dam 
was  built  up  in  successive  sections  (also  of  timber  or  stiffened  steel), 
the  horizontal  joints  being  made  by  angles  on  the  inside,  and  the  walls 
being  braced  by  transverse  struts,  where  the  shape  and  size  demanded 
it.  The  shaft  was  of  steel  tubing  fastened  to  the  roof  and  at  the  sev- 
eral horizontal  joints  by  outside  angles. 


Note. — These  papers  are  Issued  before  the  date  set  for  presentation  and  discussion. 
Correspondence  is  invited  from  those  who  cannot  be  present  at  the  meeting,  and  may  bo 
sent  by  mail  to  the  Secretary.  Discussion,  either  oral  or  written,  will  be  published 
in  a  subsequent  number  of  Proceedings,  and,  when  finally  closed,  the  papers,  with 
discussion  in  full,  will  be  published  in  Transactions. 


Papers.]  PNEUMATIC  FOUNDATIONS  FOR  BUILDINGS  213 

As  the  structure  was  sunk,  to  bring  the  upper  edge  of  each  section 
of  the  coffer-dam  near  the  ground  level,  a  new  section  of  coffer-dam 
and  a  new  section  of  shafting  were  added,  and  the  space  between  the 
coffer-dam  and  the  shafting  was  filled  with  concrete.  When  bed-rock 
was  reached,  the  working  chamber  and  the  shaft  were  also  filled  with 
concrete.  The  finished  pier  consisted  of  two  entirely  separate  bodies 
of  concrete — an  inverted  T-shaped  portion  bounded  by  the  shafting 
and  the  roof  and  walls  of  the  working  chamber,  and  a  ring-shaped  por- 
tion surrounding  the  shaft  and  enclosed  within  the  coffer-dam. 

The  surrounding  shell,  consisting  of  the  coffer-dam  and  the  sides 
of  the  working  chamber,  whether  of  timber  or  of  steel,  could  only  be 
considered  a  mould  for  the  concrete  and  a  curb  or  lining  for  holding 
back  the  earth  during  the  sinking  of  the  caisson.  It  could  not  be  cal- 
culated as  supporting  any  weight,  but,  on  the  contrary,  was  certain  to 
rot  or  corrode  in  time,  and  leave  a  more  or  less  free  space  around  the 
pier.  The  shafting,  and  especially  the  roof,  where  the  latter  was  of 
metal  and  was  left  in  place,  presented  very  serious  possibilities.  Their 
protection  from  corrosion  depended  on  the  care  with  which  the  con- 
crete was  rammed  into  contact  with  them.  If  either  corroded  to  a  sub- 
stantial extent,  it  would  produce  a  very  large  surface  of  weakness. 
The  permanence  of  these  important  elements  of  the  structure,  there- 
fore, depended  on  the  care  of  workmen,  who  are  not  to  be  relied  on  for 
more  care  than  is  necessary  at  the  moment.  Furthermore,  the  angles 
at  the  several  horizontal  joints  formed  grooves  in  the  concrete  from 
3  to  6  in.  deep.  Only  under  unusually  favorable  conditions  could  the 
shafting  angles  be  calculated  to  act  as  supporting  a  share  of  the  load 
in  the  ratio  of  their  horizontal  area  to  that  of  the  complete  cross- 
section  of  the  pier;  but  the  angles  at  the  joints  of  the  coffer-dam 
would  not  transmit  any  substantial  pressure  to  the  concrete  below 
them,  because  the  concrete  would  never  be  rammed  under  them  suffi- 
ciently. The  only  transmission  of  pressure  would  be  to  the  decaying 
or  corroding  walls,  and  the  angles  themselves  would  corrode  in  time. 
The  greatest  area  upon  which  the  bearing  strain  could  be  calculated 
correctly,  therefore,  was  that  within  the  inner  edge  of  the  angle-irons 
(X,  Fig.  2),  rather  than  that  within  the  inner  face  of  the  coffer-dam 
{Y,  Fig.  2).  As  a  matter  of  fact,  the  latter  standard  was  generally 
used,  but  the  error  was  swallowed  in  the  large  factor  of  safety  made 
necessary  by  the  uncertainties  of  the  problem. 


214  PNEUMATIC  FOUNDATIONS  FOR  BUILDINGS  [Papers. 

furthermore,  the  useless,  and  to  some  extent  harmful,  materials 
left  in  the  ground,  were  very  expensive  parts  of  the  structure. 

There  were  thus  two  powerful  incentives  for  the  elimination  of 
these  materials  from  the  finished  structure,  either  by  sinking  the  pier 
without  them,  or  by  withdrawing  them  after  use.  ISTevertheless,  there 
was  a  period  of  many  years  during  which  little  or  nothing  was  accom- 
plished. 

The  recent  activity  in  high  building  construction  in  New  York 
City,  however,  making  necessary  a  very  extensive  use  of  caissons  of 
this  type,  has  witnessed  the  substantial  elimination  of  every  material 
but  concrete.  The  sinking  of  the  coffer-dara  and  of  a  metal  or  timber 
roof  for  the  working  chamber,  has  been  rendered  unnecessary,  and  the 
steel  shafting  has  been  designed  to  permit  its  ready  removal  after  it 
has  served  its  purpose  in  the  sinking  of  the  caisson.  These  improve- 
ments have  been  put  into  practice  in  the  foundations  of  the  building 
for  the  United  States  Express  Company,  at  the  corner  of  Rector  Street 
and  Trinity  Place;  the  New  Trinity  Building;  the  building  for  the 
United  States  Realty  Company,  at  Broadway  and  Thames  Street,  and 
the  Singer  Building,  on  Broadway  near  Liberty  Street. 

Elimination  of  the  Roof. — The  most  serious  objection  to  caissons 
of  the  style  described  has  been  the  existence  of  the  roof,  constituting 
a  dividing  plane  across  almost  the  entire  cross-section.  The  objection 
to  such  a  dividing  plane  was  appreciated  from  the  earliest  use  of 
pneumatic  caissons.  The  late  Theophilus  E.  Sickles,  M.  Am.  Soc.  C. 
E.,  in  1870,  and  John  F.  O'Rourke,  M.  Am.  Soc.  C.  E.,  in  1898,  pro- 
posed the  removal  of  the  roof  after  the  sinking  of  the  caisson  and  be- 
fore the  introduction  of  the  concrete  above  the  working  chamber. 

The  Sickles  caisson  is  shown  in  Eigs.  3  and  4.  The  roof  consisted 
of  four  segmental  plates  bolted  to  the  under  side  of  internal  flanges 
of  the  casing  and  attached  to  each  other  by  bolts  passing  through 
radial  flanges  on  the  under  side.  After  sinking  to  the  required  depth, 
and  sealing  the  cutting  edge  with  a  sufficient  filling  of  concrete  to  pre- 
vent the  entrance  of  water,  the  air  was  cut  ofl^  and  the  roof  removed 
by  withdrawing  the  bolts  passing  through  the  several  flanges.  The 
caisson  of  the  type  shown  had  a  high  roof  and  no  separate  air-shaft 
siapported  upon  the  roof,  as  in  the  modern  type,  the  cofl'er-dam  or  outer 
shell  being  made  air-tight  throughout  its  height.  For  a  caisson  of  this 
type,  the  design  of  the  roof  was  probably  entirely  satisfactory. 


Pii  IK- 


PNEUMATIC    I'OUXDATIOXS   FOIt    BUll.D[N(iS 


215 


The  O'Rourke  caisson.  Figs.  5  and  6,  utilized  a  similar  roof  in 
lialf-roiind  sections,  but  the  roof  was  bolted  on  the  top  of  the  inward 
flange  of  the  casing,  and  the  flanges  connecting  the  segments  to  each 
other  were  at  the  top.  This  would  permit  the  filling  of  the  working 
chamber  with  concrete  clear  up  to  the  roof  before  removing  the  latter. 


Fig.  6 


Fig 


Fig,  8 


Fig.  9 


Fig.  10 


The  chief  defect  of  these  methods,  however,  appears  in  cases  where, 
in  order  to  get  the  requisite  weight,  the  concrete  is  filled  into  the  space 
above  the  roof  during  the  sinking  operation,  as  is  usual  in  sinking 


216  PNEUMATIC   rOUXDATIOXS  FOR  BUILDINGS  [Papers. 

through  earth  for  building  foundations.  In  such  operations  it  ha3 
been  impossible  to  eliminate  the  roof  of  the  working  chamber  until 
the  introduction  of  a  recent  improvement  which,  at  the  same  stroke, 
eliminated  the  coffer-dam  which  had  previously  passed  for  a  necessary 
evil  in  sinking  caissons  in  earth.  The  feasibility  of  the  improvement 
was  first  demonstrated  by  sinking  all  the  caissons  for  the  building  for 
the  United  States  Express  Company  by  this  method,  and  at  a  sub- 
stantial reduction  in  cost. 

Elimination  of  the  Coffer-dam. — There  had  been  previously  sug- 
gested, in  1904,  the  elimination  of  the  coffer-dam  and  roof  by  sinking 
practically  a  solid  pier  of  concrete,  with  only  a  central  air-shaft  and 
a  working  chamber  hollowed  out  of  the  bottom.  Fig.  7  gives  a  suffi- 
cient idea  of  the  construction  proposed.  There  was  no  distinction  be- 
tween different  parts  of  the  structvire,  except  in  so  far  as  the  lower 
portion  of  the  concrete  might  be  considered  as  the  roof  and  side  walls 
of  the  working  chamber,  and  the  concrete  above  this  might  be  con- 
sidered as  the  coffer-dam  extending  solidly  from  the  surrounding 
earth  to  the  shaft.  It  was  proposed  to  build  the  whole  of  annular 
blocks  of  concrete  laid  one  above  another,  or  to  form  substantially  a 
monolith  by  building  up  the  structure  in  situ  as  fast  as  it  was  sunk. 
The  difficulties  in  the  way  of  moulding  the  concrete  working  chamber 
with  suitably  strong  roof  and  sides  and  hardening  it  sufficiently  in  the 
short  time  available  at  the  works  then  in  hand  prevented  the  utiliza- 
tion of  this  design,  and,  instead,  the  contractors  adopted  the  design 
shown  in  Figs.  8,  9,  and  10. 

The  working  chamber  was  built  of  heavy  timber,  and  across  the 
top  were  laid  angle-irons,  a  few  inches  below  which  was  fastened  a 
temporary  flooring.  The  steel  shafting  was  supported  on  this  flooring, 
and  a  roof  of  concrete  was  moulded  thereon  to  a  substantial  height, 
and  of  the  same  outside  dimensions  as  the  working  chamber.  The 
earth  being  excavated,  and  the  chamber  sunk  to  a  sufficient  depth, 
another  section  of  concrete  was  added.  The  shafting  was  built  up 
from  time  to  time  to  maintain  it  above  the  concrete.  After  the  first 
section  of  concrete  was  finished,  the  successive  sections  were  moulded 
in  place  without  interruption  of  the  sinking  operations;  the  excava- 
tion and  the  building  up  proceeding  of  course  at  the  same  ultimate 
rate,  but  quite  independently  of  ench  other,  and  the  coffer-dam,  reduced 
to  merely  a  mould  for  the  concrete,  being  removed  before  the  sinking 
of  each  concrete  section. 


Papers.]  PNEUMATIC   FOUNDATIONS  FOR   BUILDINGS  217 

In  a  previous  design,  it  had  been  proposed  to  divide  each  section 
of  the  coffer-dam  into  flat  nnits  which  might  be  readily  transported 
and  only  united  to  each  other  when  in  place  on  the  next  lower  section, 
this  method  having  the  further  advantage  of  avoiding  the  necessity  of 
breaking  the  air-pipes  (see  Fig.  1),  which  had  been  a  cause  of  delay 
with  the  use  of  sections  which  were  completed  before  being  put  in 
place;  and  such  flat  units  were  now  found  to  be  excellent  moulding 
plates,  only  four  being  needed  for  each  section  of  concrete,  and  ex- 
cessive lengths  being  unobjectionable,  because  one  might  overlap  the 
next  at  the  corner. 

The  temporary  flooring  carried  the  concrete  roof  until  the  latter 
was  hardened,  and  was  removed  before  putting  on  the  air  pressure  and 
the  necessary  lock.  The  angle  cross-bars  remained  embedded  in  the 
concrete,  transmitting  its  weight  to  the  timber  walls,  although  they 
were  not  necessary  for  the  purpose  after  the  concrete  had  hardened; 
and,  in  fact,  after  reaching  a  comparatively  slight  depth,  the  weight 
of  the  concrete  was  sustained  by  the  skin  friction  and  the  air  pres- 
sure, and  added  weights  were  necessary  to  force  the  caisson  down. 
The  cross-bars  might  have  been  designed  and  connected  so  as  to  per- 
mit their  removal  after  the  hardening  of  the  concrete,  if  such  removal 
had  been  thought  of  importance. 

Only  one  accident  occurred,  and  this  demonstrated  the  advisability 
of  using  timber  rather  than  concrete  for  the  walls  of  the  working 
chamber.  The  earth  under  one  wall  of  the  working  chamber  had  been 
excavated  previously  to  remove  the  footing  of  an  old  wall.  When  the 
first  section  of  concrete  had  been  moulded  on  this  working  chamber  and 
the  mould  had  been  removed,  preparatory  to  sinking  the  concrete  sec- 
tion, the  old  material  replaced  in  the  excavation  allowed  one  side  to 
settle  so  as  to  tilt  the  structure,  and,  before  it  could  be  righted,  it  fell 
over.  The  concrete  was  tied  to  the  working  chamber  only  by  the  cross- 
ing angles  embedded  in  the  base  of  the  concrete,  and  swung  bodily 
about  the  upper  edge  of  a  side  wall  of  the  working  chamber,  thus  for 
a  time  putting  its  entire  weight  on  this  single  wall.  But  the  chamber 
was  built  so  strongly  that  it  was  substantially  uninjured,  and  the 
workmen  in  it  at  the  time  were  unscathed.  The  accident,  while  in- 
dicating the  necessity  for  greater  precaution  in  building  and  sinking 
the  first  concrete  section,  demonstrated  the  practical  excellence  of  the 
design. 


218  PNEUMATIC   FOUNDATIONS  FOR  BUILDINGS  [Papers. 

When  such  a  caisson  was  sunk  to  its  final  depth,  there  was  no  metal 
or  timber  roof  to  be  removed.  The  cost  of  making  first  a  sectional 
bolted  roof,  like  that  of  Sickles  or  O'Rourke,  and  subsequently  remov- 
ing it,  was  saved;  and,  which  is  probably  more  important,  the  intro- 
duction of  concrete  above  the  working  chamber  did  not  have  to  await 
the  sinking  of  the  caisson.  Its  weight  could  be  utilized  in  the  sink- 
ing of  the  structure,  and  this  weight,  in  caissons  passing  for  a  great 
depth  through  earth,  is  a  very  substantial  consideration.  It  consti- 
tuted probably  the  greatest  of  the  series  of  advance  steps  under  dis- 
cussion. 

Elimination  of  Shaft  Lining. — The  finished  pier  included,  besides 
the  concrete  body,  the  cross-bars,  which  are  a  negligible  consideration, 
being  entirely  embedded  so  as  to  prevent  corrosion,  and  being  of  such 
slight  cross-section  as  not  to  form  cleavage  planes  in  the  concrete; 
and  the  steel  shaft  lining,  which,  at  the  very  best,  added  not  a  pound 
to  the  load  for  which  the  pier  might  be  safely  designed,  and,  at  the 
worst,  might  prove  an  element  of  weakness,  and  was  certainly  an  ele- 
ment of  substantial  expense. 

The  progress  of  improvement  in  eliminating  the  shaft  lining  was 
the  reverse  of  that  in  eliminating  the  roof.  In  the  latter  case,  the 
idea  was  first  advanced  of  making  the  roof  removable  after  the  caisson 
had  been  sunk;  and  the  successful  solution  of  the  problem  lay  in  avoid- 
ing the  building  of  a  true  roof.  In  the  case  of  the  shaft  lining,  the  first 
proposals  endeavored  to  avoid  its  use  entirely,  but  practical  success 
came  only  with  the  idea  of  sinking  the  caisson  with  a  shaft  lining 
similar  to  those  previously  used,  and  removing  the  lining  after  sink- 
ing and  before  introducing  the  filling  of  concrete. 

The  first  idea  is  shown  in  Fig.  11.  A  shaft  lining  of  moulded  con- 
crete is  shown.  To  avoid  excessive  loss  by  leakage  of  air  through  the 
concrete,  it  was  proposed  to  coat  the  inner  surface  of  the  shaft  lining 
with  air-tight  material,  such  as  a  paint  containing  lime.  The  difficulty 
of  connecting  the  shaft  lining  to  the  air-lock  with  sufficient  strength  to 
resist  the  upward  air  pressure  on  the  latter  was  to  be  obviated  by  long 
tie-rods  extending  from  the  lock  to  the  lowest  section  of  the  shaft 
lining,  as  indicated  in  dotted  lines.  It  was  also  proposed  in  this  de- 
sign to  eliminate  the  lining  entirely,  merely  coring  the  concrete  body 
and  coating  the  surface  with  paint,  as  above,  the  manner  of  fastening 
the  air-lock  not  being  specified. 


Papers.]  PNEUMATIC  FOUNDATIONS  FOR  BUILDINGS 


219 


^%wj!,\>«te^ 


220  PNEUMATIC   FOUNDATIONS  FOR  BUILDINGS  [Papers. 

The  first  successful  attempt  to  eliminate  the  shaft  lining,  however, 
involved  the  i;se  of  a  removable  lining,  which,  while  costing  more  than 
those  of  common  design,  is  usable  again  and  again  indefinitely,  and, 
in  the  long  run,  effects  a  great  economy.  The  design  used  in  sinking 
the  caissons  of  the  new  Trinity  addition,  and  the  adjoining  bviilding 
of  the  United  States  Realty  Company,  is  shown  in  Figs.  12,  13,  14, 
and  15.  It  was  found  that  a  comparatively  small  number  of  sections 
served  for  the  sinking  of  many  piers.  There  was  no  material  loss  of 
time  involved  in  removing  the  sections  and  reassembling  them  for  fur- 
ther use.  In  fact,  the  job  was  completed  in  much  less  than  the  previous 
record  time  for  such  work. 

Figs.  12  and  13  show  the  shaft  lining  in  place;  Figs.  11:  and  15 
show  the  construction  of  one  of  the  collapsible  sections.  Each  section 
was  composed  of  two  approximately  semicircular  plates  internally 
flanged  for  bolting  to  each  other  along  one  vertical  edge,  and  a  key 
interposed  between  the  opposite  edges  of  the  plates.  Internal  flanges 
at  the  ends  served  for  bolting  successive  sections  to  each  other.  Ladder 
rungs  were  arranged  conveniently  between  the  flanges  of  the  key,  and 
vertical  guides  were  arranged  just  inside  the  line  of  the  end  flanges  to 
guide  the  bucket  past  them.  In  some  cases  the  tubing  was  made  oblong 
in  cross-section  instead  of  circular.  Packing  was  provided  in  all  the 
joints,  and  this  was  the  only  part  of  the  striicture  requiring  renewal, 
it  being  cheaper  to  provide  new  packing  for  each  re-use  than  to  try  to 
save  the  old. 

Fig.  16  shows  the  finished  pier,  supposing  the  working  chamber  to 
l>e  built  of  sheet  steel.  The  dotted  line  indicates  the  joint  between 
the  concrete  set  up  in  sinking  the  pier  and  the  filling  introduced  after- 
ward. 

Comparison  with  Concrete  Piles. — Side  by  side  with  the  lorogress 
in  caisson  work,  recent  years  have  seen  a  rapid  inipnnement  in  the 
sinking  or  building  of  concrete  piles  in  the  earth.  The  first  attempts 
to  substitute  concrete  for  timber  or  steel  in  piles  contemplated  the 
manufacture  of  the  concrete  piles  above  ground  and  the  sinking  of 
them  by  one  or  another  of  the  methods  used  for  timber  or  steel  piles. 
But,  at  present,  there  are  in  the  market  several  styles  of  concrete  piles 
made  by  first  forming  the  excavation  and  subsequently  filling  in  the 
concrete.  These  methods  permit  the  formation  of  piles  of  great  depth 
and  of  theoretically  unlimited  diameter.     Starting  from  widely-sepa- 


Papers.]  PNEUMATIC   FOUNDATIONS  FOR  BUILDINGS  221 

rated  points,  the  two  arts,  caisson  work  and  pile  work,  have  constantly 
converged  toward  the  same  goal,  a  simple  concrete  column,  bearing 
upon  a  rock  or  similar  sub-foundation  in  the  case  of  caissons  and  some 
piles,  and  supported  by  skin  friction  in  the  case  of  other  piles. 

The  analogy  has  been  carried  even  further  by  more  recent  improve- 
ments in  which  vertical  reinforcing  rods  of  steel,  similar  to  those 
sometimes  used  in  concrete  piles,  are  embedded  in  the  concrete  of  the 
pier.  The  base  of  such  a  pier  is  shown  in  vertical  section  in  Fig.  17, 
and  Fig.  18  shows  a  concrete  pile  similarly  reinforced.  The  reinforcing 
rods  in  the  pier  should  extend  down  to  the  rock  sub-foundation,  and 
are  most  easily  introduced  in  that  method  of  construction  in  which  the 
rcof  of  the  working  chamber  is  omitted,  turn-buckles  being  introduced 
for  putting  the  rods  under  stress  before  embedding  them  in  concrete. 
The  non-adjustable  flange  joints  may  be  used  for  the  rods  which  run 
through  the  shaft,  and  substantially  the  entire  length  of  which  may 
bear  freely  on  the  sub-foundation  before  the  concrete  is  filled  in  about 
them. 

Most  Becent  Modifications. — The  steel  rods  in  the  foregoing  de- 
signs merely  reinforce  the  concrete.  Should  the  concrete  fail,  or  be 
designed  or  built  so  as  to  shift  a  substantial  portion  of  the  load  to  the 
rods,  the  latter  would  be  unable  to  stand  the  strain.  A  recent  design 
includes  the  introduction  of  columns  of  sufficient  strength  to  carry  a 
substantial  load.  In  fact,  they  may  be  proportioned  to  carry  all  or 
the  greater  part  of  the  load.  Fig.  19  shows  the  caisson  sunk  to  rock, 
and  the  columns  in  place,  ready  to  be  filled  with  concrete.  The  columns 
are  of  ordinary  style,  built  up  of  Z'^bars  riveted  to  a  central  plate.  One 
column  is  embedded  in  the  concrete  from  the  beginning,  and  is  wedged 
up  at  its  lower  end.  This  column  may  be  duplicated  as  often  as  de- 
sired. Another  passes  down  through  the  shaft,  and  is  properly  sup- 
ported before  its  embedment.  The  shaft  lining  may  or  may  not  be 
withdrawn,  as  desired. 

Since  it  is  possible  to  carry  concrete  piles  in  many  cases  to  a  rock 
sub-foundation,  where  they  act  as  true  cohimns,  the  idea  has  been  con- 
ceived of  substituting  steel,  with  its  immensely  greater  strength  as  a 
column,  and  surrounding  it  with  concrete,  which  stiffens  the  column 
to  some  extent,  but  which  performs  the  principal  function  of  protect- 
ing the  steel  from  corrosion.  The  finished  pile  or  column  is  indicated 
in  Figs.  20  and  21.     The  column  is  hollow,  which  serves  to  carry  a 


222  PNEUMATIC   FOUNDATIONS  FOR  BUILDINGS  [Papers. 

water-jet  for  sinking  the  column  itself,  and  has  a  surrounding  shell, 
which  is  afterward  filled  with  concrete  around  and  within  the  center 
of  the  column.  The  shell  may  be  withdrawn  as  the  concrete  is  intro- 
duced. The  column  may  be  shod  at  its  lower  end  so  as  to  secure  a 
good  bearing  by  ramming  it  down  on  the  rock. 

Invention  is  largely  accidental,  and  its  progress  is  apt  to  be  most 
erratic.  The  writer  has  never  observed  a  series  of  improvements  pro- 
gressing more  logically  and  consistently  in  the  same  direction  than 
those  here  considered.  The  engineering  profession  owes  to  Daniel  E. 
Moran,  M.  Am.  Soc.  C.  E.,  and  John  W.  Doty,  Assoc.  M.  Am.  Soc.  C. 
E.,  who  conceived  these  improvements,  and  to  the  Foundation  Com- 
pany, by  whom  they  were  put  into  practice,  a  very  large  debt  for  the 
originality  and  progressive  spirit  with  which  they  have  met  the  de- 
mands of  modern  builders  for  economical  methods  of  providing  founda- 
tions of  maximum  bearing  strength. 


Vol.  XXXIV.  MARCH,   1908.  No.  3. 


AMEEICAN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED    185  2. 


PAPERS  AND   DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


SUBSTRUCTURE   OF   PISCATAQUIS   BRIDGE, 
AND  ANALYSIS  OF  CONCRETE  WORK. 


By  G.  A.  Hersey,  Jun.  Am,  Soc.  C.  E. 
To  BE  Presented  May  6th,  1908. 


This  paper  gives  a  general  description  of  the  construction  of  the 
Piscataquis  Kiver  Bridge,  built  for  the  Bangor  and  Aroostook  Kail- 
road,  and  also  the  results  attained  with  the  different  classes  of  con- 
crete used. 

The  Piscataquis  Bridge  was  built  during  1907,  and  is  a  part  of  the 
"Medford  Cut-off,"  an  extension  of  the  Northern  Maine  Seaport  Kail- 
road,  a  branch  of  the  Bangor  and  Aroostook  Kailroad.  This  extension 
begins  at  the  present  terminus  of  the  Northern  Maine  Seaport  Kail- 
road,  and  runs  northward- about  28_miles  until  it  again  strikes  the 
main  line  of  the  Bangor  and  Aroostook  Kailroad.  It  shortens  the 
distance  between  the  two  points  on  the  main  line  4.3  miles,  reduces 
the  curvature  considerably,  and  gives  much  easier  grades. 

The  "Cut-off"  crosses  the  Piscataquis  Kiver  in  the  Town  of  Med- 
ford, and  on  the  line  of  a  very  high  horse-back — a  formation  peculiar 
to  that  section  of  the  country — which  was  of  considerable  value  in  the 
construction  of  the  railroad.  The  line  follows  the  horse-back  in  a  gen- 
eral direction  for  about  14  miles,  and  for  6  miles  skirts  along  its  side; 
it  can  even  be  said  that  the  entire  road  was  made  from  it,  for,  as  the 

Note. — These  papers  are  issued  before  the  date  set  for  presentation  and  discussion. 
Correspondence  is  invited  from  those  who  cannot  be  present  at  the  meeting,  and  may  be 
sent  by  mail  to  the  Secretary.  Discussion,  either  oral  or  written,  will  be  published 
in  a  subsequent  number  of  Proceedinps.  and,  when  finally  closed,  the  papers,  with 
discussion  in  full,  will  be  published  in  Transactions. 


224  BRIDGE  SUBSTRUCTURE  [Papers. 

northerly  half  of  the  line  passes  through  low  land,  material  from  the 
horse-back  was  used  for  filling,  as  well  as  ballasting.  The  material  in 
it  varies  from  sand  to  coarse  gravel,  and,  in  a  few  instances,  clay.  At 
the  river  the  best  kind  of  gravel  was  found,  and  the  hills  on  either 
side  afforded  excellent  sand  and  stone  for  concrete. 

The  grade  of  the  railroad  is  55.5  ft.  above  the  average  water  level, 
with  about  8  ft.  of  water  in  the  river. 

A  bridge  of  the  deck  type  was  adopted,  with  four  river  piers  and 
two  shore  abutments  of  reinforced  concrete.  The  line  crosses  at  a 
bend  in  the  river,  the  piers  being  placed  at  an  angle  of  55  degrees. 
The  total  length,  of  the  bridge  is  607  ft.  10  in.  About  13  tons  of  steel 
were  used  for  reinforcement,  mostly  in  the  two  abutments,  there  be- 
ing but  little  placed  in  the  tops  of  each  of  the  piers.  > 

The  work  was  handled  with  a  Lidgerwood  cableway,  800  ft.  long, 
placed  on  the  center  line  of  the  bridge.  This  cableway  was  used  in 
making  all  the  excavation,  in  conveying  and  placing  concrete,  moving 
machinery,  and,  later,  in  erecting  the  temporary  trestle  bridge.  The 
cableway  clearly  demonstrated  its  suitability  in  this  case,  and,  for 
rapid  and  profitable  work,  it  would  be  hard  to  find  anything  better. 
In  landing  the  north  abutment,  it  was  necessary  to  go  about  50  ft. 
into  the  side  of  a  40-ft.  bank  and  remove  about  3  000  cu.  yd.  With 
the  cableway,  all  this  material  was  saved  and  used  directly  for  con- 
crete and  for  banking  coffer-dams,  whereas,  by  almost  any  other  method, 
it  would  have  been  necessary  to  rehandle  it  several  times. 

The  concrete  was  all  machine-mixed,  and  dumped  into  buckets 
which  were  run  out  under  the  cableway  and  carried  to  any  part  of  the 
bridge.  The  greatest  number  of  buckets  used  in  one  day  was  182,  for 
9  hours'  work.    Each  bucket  held  1  cu.  yd. 

Crib  coffer-dams,  of  8  by  8-in.  timber,  in.  8-ft.  sections,  were  made 
on  the  river  bank.  Alternate  sections  were  floored  about  four  tiers  from 
the  bottom.  These  cribs  were  then  set  in  place,  and  the  floored  sections 
were  loaded  with  rock.  The  outside  was  covered  with  2-in.  planks 
drive'n  into  the  river  bottom  as  far  as  possible  by  hand-mauls.  The 
cribs  were  then  banked  with  earth  to  above  the  water  level.  These 
coffer-dams  gave  excellent  satisfaction,  and  only  in  one  instance  was 
there  any  trouble  from  leakage,  and  that  was  quickly  remedied  by  a 
generous  use  of  straw  and  gravel.  The  pumping  was  done  by  five 
centrifugal  pumps  having  a  combined  discharge  of  24  in.,  and  they 
were  able  at  all  times  to  take  care  of  the  water. 


PLATE  XXXII. 
PAPERS,   AM.  SOC.   C.   E. 

MARCH,   1908. 

HERSEY  ON 

BRIDGE  SUBSTRUCTURE. 


Fig.  1.— South  Abutment,  Showing  Method  of  Placing  Concrete  in  Finishing 
Abutment.    Total  Height,  57  Ft.     (Above  Ground,  45  Ft.) 


Fig.  2.— Building  Coffer-Dam  of  Pier  3,  Showing  Cableway. 


l*«P<'i'«]  BRIDGE  SUBSTRUCTUEE  225 

The  river  bottom  is  very  rocky  and  gravelly,  and  almost  as  hard 
and  compact  as  if  cemented.  This  feature  could  not  have  been  im- 
proved upon  for  the  foundation.  Excavations  vpere  carried  down  to 
an  average  depth  of  about  5  ft.  below  this  bottom,  and  all  footings 
rested  on  hard  gravel  through  which  test-bars  could  not  be  driven  more 
than  2  ft.  The  piers  were  liberally  rip-rapped  with  the  rock  which 
had  been  used  to  load  tlie  coffer-dams,  and  other  larger  rock. 

Two  mixtures  of  concrete  were  used,  namely :  1:2:4  for  under 
water,  and  1:3:5  for  above  water.  Suitable  gravel  was  found  mixed 
with  sand  in  about  the  right  proportions  and  was  used  without  screen- 
ing. Daily  tests  of  the  aggregates  were  made,  by  volumes,  so  that  it 
was  known  that  the  proper  ratios  were  being  maintained.  The  sand 
in  the  gravel  was  very  clean  and  sharp,  and  free  from  loam  and  clay. 
The  treatment  of  the  aggregates  was  as  follows:  The  specification 
called  for  measurement  by  volume,  1  bbl.  of  cement,  3  bbl.  of  sand,  and 
5  bbl.  of  stone,  etc.  Test  boxes,  holding  half  a  batch  of  gravel,  were 
filled  with  the  aggregates,  as  placed  upon  the  mixing  platform,  the 
contents  were  screened,  and  the  sand  and  rock  measured  separately. 
If  not  in  the  right  proportions  more  sand  or  more  rock  was  added,  as 
the  case  required.  By  making  daily  tests,  and  inspecting  closely  the 
materials  as  used,  the  proper  proportions  were  maintained. 

Table  1  shows  the  results  obtained,  and  it  is  interesting  to  compare 
them  with  the  results  from  mixtures  made  under  ideal  conditions,  or 
those  oblained  where  the  ingredients  were  screened  and  graded  more 
carefully.     The  ratios  were  determined  by  weight  as  well  as  by  volume. 

In  making  the  calculations  in  Table  1  for  the  number  of  cubic 
feet  of  concrete  per  barrel  of  cement,  the  quantities  used  in  the  putty 
coats  were  not  considered,  as  so  few  barrels  were  used  that  they  would 
not  have  much  effect  on  the  results.  The  1:2:4  mixture  took  1.36 
bbl.,  and  the  1:3:5  mixture,  1.21  bbl.  of  cement  per  cubic  yard  of 
concrete.  A  perfect  mixture  of  the  1:2:4  class  would  require  1.46 
bbl.  per  cu.  yd.,  and  of  the  1:3:5  class,  1.11  bbl.  per  cu.  yd.,  which, 
as  compared  with  the  results  obtained,  shows  that  the  1:2:4  mixture 
fell  short  x&  l^^bl.  per  cu.  yd.,  and  the  1:3:5  mixture  over-ran  ^o 
bbl.  per  cu.  yd.  This  would  indicate  that  the  cement  used  in  the  entire 
bridge  was  83.88  bbl.  more  than  called  for  theoretically. 

The  fact  that  the  1:2:4  mixture  is  short  in  cement  and  that  the 
1:3:5  mixture  has  a  surplus,  may  be  accounted  for  in  two  different 


226 


BRIDGE  SUBSTRUCTURE 


[Papers. 


O  o  o  o  o  o 

g.  i(i.  CO  10  l-i  ^ 

!>  3  2  3  tS  >>» 

'  -!  "-i  r-  "■  C 


O  UTOD0D«D 
O  -1  CO  00  CO 


l-"  -J  tc  CO  to 
COOI-'CCCO 
i-iiOi-1-J-J 


ooooo 
ooooo 


-JoaoCOCn 
ooooo 


^t^^a^s^o^ 


OM  CDOSCD  O 
03  «».  i-J.  05  to  O 


CJI  CO  CO  CO  to  bS 
O  O  O;.  *^  GC  0> 
CO  JO  0»  d  O  1-' 

ot'o'oo'oo» 

oooooo 


tOiO  JO  JOtOlO 
JO  I-'  M- JO  JO  JO 

bo  it^  'c»  ~j  o  o 
ooooo  o 


cei-icoco  eo>c> 

o  'c  o  o  o  o 


to  &"  CO  i-i  o 
to  CO  >-•  hJ  !-■  ox 

'o  00  -J  bi  to  o 

CO-i  iti.  1-'  JO  o 


OS  »^  oi  oi  m  JO 
1-' oo  oo  00  JO  en 
jocnoooocn 

en  en  o  o  o  ci» 
oooooo 


£cs. 


3  -i  trj 


o 
^  »  « 


o  3 


^  C  r? 


bd 


w 

H 

cc 

O 

td 

;> 

I— I 


a 

Q 
O 

H 


o 
w 

W 
*^ 

o 

w 

a 

M 

;» 
O 

,© 

d 


W 


td 
w 

5 
o 


PLATE  XXXIII. 

PAPERS,   AM.  SOC.   C.   E. 

MARCH,   1908. 

HERSEY  ON 

BRIDGE  SUBSTRUCTURE. 


Fig.  1.— General  View.  Locking  South.    North  Abutment  and  Piers  1  and 
Finished.  Pier  S  Nearly  Complete,  and  Pier  4  Being  Excavated. 


Fig.  3.— General  View  of  Completed  Bridge. 


Papers.]  BRIDGE  SUBSTRUCTURE  237 

ways :  First,  the  aggregates  were  not  graded  properly,  so  that  the  voids 
were  not  completely  filled;  second,  the  quantities  of  the  aggregates 
used  for  the  different  mixtures  were  not  always  exactly  the  same.  The 
material  was  loaded  into  the  mixer  by  wheel-barrows,  the  necessary 
quantity  for  a  wlieel-barrow  for  each  mixture  was  determined,  and 
they  were  all  supposed  to  be  loaded  the  same;  in  this,  however,  there 
was  bound  to  be  some  variation. 

For  1  bbl.  of  cement,  or  4  bags  (all  the  cement  being  in  bags)  8 
wheel-barrows  of  the  aggregate  were  used  in  the  1:3:5  mixture,  and 
6  wheel-barrows  of  the  aggregate  in  the  1:2:4  mixture.  A  check 
was  had  on  the  quantities  used  by  the  space  occupied  by  the  completed 
batch  in  the  bucket,  after  being  dumped  from  the  mixer,  as  the 
buckets  each  held  1  cu.  yd.,  it  was  found  by  test  samples  just  how  full 
they  should  be  for  the  two  different  mixtures.  Then,  again,  the  ag- 
gregates not  being  uniformly  graded  by  Nature  would  have  a  tendency 
to  throw  the  resultant,  cubic  feet  per  barrel,  under  or  above  the 
theoretical  results  obtained  from  perfect  mixtures.  It  seems,  however, 
that  where  suitable  ingredients  can  be  found  in  their  natural  state, 
free  from  loam  and  clay — although  a  small  percentage  of  either  will 
not  decrease  the  strength  of  the  concrete — that  as  good  work  at  less 
cost  per  yard  can  be  obtained  as  where  the  sand  is  screened  from  the 
gravel  and  they  are  again  mixed  artificially;  for,  when  the  ingredients 
go  into  the  mixer  in  their  natural  state,  the  machine  has  to  do  less 
work  in  mixing  the  cement  with  them,  than  when  the  sand,  rock,  and 
cement  go  in  separately,  for  the  mixer  has  not  only  to  mix  the  cement 
in,  but  the  sand  and  rock  as  well.  Thus,  with  a  mixer  running  for 
the  same  time,  under  the  two  different  conditions,  it  would  seem  that 
a  better  mixture  could  be  made  from  the  natural  ingredients.  There 
were  instances  during  the  progress  of  the  work  when  the  resultant 
mixtures  agreed  exactly  with  the  required  quantity  of  cement  per  barrel. 
This  would  indicate  that  in  such  cases  the  voids  were  completely  filled, 
the  sand  filling  the  interstices  of  the  rock  and  the  cement  those  of 
the  sand. 

The  piers  were  designed  with  round  noses,  having  a  slight  batter. 
At  a  point  well  above  high  water,  on  the  front  ends,  a  circular  break- 
water was  made,  having  a  batter  of  2  :  3  and  extending  down  to  with- 
in 3  ft.  of  the  footing  course.  The  forms  for  these  noses  were  made 
of  2-in.  plank,  about  4  in.  wide,  sections  being  built  up  in  8-ft.  lengths 


228  BRIDGE  SUBSTRUCTURE  .  [Papers. 

as  the  concrete  advanced.  With  plank  of  this  width,  the  circular 
form  could  be  made  very  readily,  and  with  very  smooth  surfaces. 

The  concrete  was  mixed  wet,  no  tamping  being  required,  other  than 
the  shoveling  over  it  received  after  being  dumped  from  the  bucket. 
Care  was  taken,  however,  that  the  sides  of  the  forms  were  well  worked 
around  with  shovels,  which  kept  the  rock  back  and  allowed  the  soft 
material  to  come  to  the  outside.  The  piers  when  stripped  and  dry 
received  a  coat  of  whitewash.  The  writer  is  not  wholly  convinced  of 
the  worth  of  this  coat  for  work  of  this  class,  as  it  usually  cracks  and 
peels  off.  Where  a  good  surface  has  been  obtained,  he  would  prefer  to 
omit  the  whitewash  coat. 

Although  the  season  was  unusually  wet,  the  progress  of  the  work 
was  delayed  only  for  a  few  days.  The  river  is  affected  rapidly  by  the 
rains,  there  being  no  storage  in  its  water-shed,  and  the  water  rises  and 
falls  quickly.  High  water  was  encountered  only  once,  when  the  sec- 
ond and  third  piers  were  first  started,  but  the  only  damage  was  the 
washing  away  of  a  little  of  the  coffer-dam  embankment. 

In  the  construction  of  the  temporary  falsework,  piles  were  driven 
from  a  driver  on  a  barrel  raft.  In  no  case  could  the  piles  be  driven 
more  than  5  ft.,  and  the  average  was  about  3  ft.  This  proved  fully 
the  firmness  of  the  entire  river  bottom,  as  found  during  the  excava- 
tion at  the  pier  locations.  Most  of  the  piles  were  furnished  with  steel 
points.  For  absolute  safety,  piles  should  not  be  driven  without  some 
protection  for  the  point,  no  matter  through  what  kind  of  ground,  as 
one  can  never  tell  what  material  a  pile  is  to  pass  through. 

The  contractor  for  the  concrete  work  was  Mr.  J.  B.  Mullen,  who  has 
had  wide  experience  in  similar  work.  Moses  Burpee,  M.  Am.  Soc.  C. 
E.,  is  Chief  Engineer  of  the  Bangor  and  Aroostook  Railroad.  W.  S. 
McEetridge,  M.  Am.  Soc.  C.  E.,  was  Engineer  of  Construction,  in 
charge  of  the  "Medford  Extension,"  and  the  writer  was  Engineer  in 
Charge  at  the  bridge. 


Vol.    XXXIV.  MARCH,    1908.  No.  3. 


AMEEIOAN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED    1852. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


OVEEHEAD    CONSTRUCTION    FOR    HIGH-TENSION 

ELECTRIC  TRACTION   OR   TRANSMISSION. 

Discussion.* 


By  IMessrs.  Joseph  Mayer,  W.  K.  Archbold,  Charles  Eufus  Harte, 
Farley  Osgood,  and  W.  S.  Murray. 


Joseph  Mayer,  M.  Am.  Soc.  C.  E. — This  paper,  in  the  main,  is  ^r  Mayer, 
an  interesting  collection  of  data  and  tables,  useful  in  the  design  of 
overhead  contact  and  transmission  lines. 

The  tables  of  wind  velocities  and  pressures  are  especially  useful 
for  forming  a  correct  opinion  of  the  actual  pressures.  More  stress 
might  be  laid  on  the  fact  that  the  observations  of  the  Weather  Bureau 
are  made  on  the  tops  of  high  buildings,  while  the  transmission  lines, 
and  especially  the  contact  lines,  are  near  the  surface  of  the  ground, 
where  the  wind  pressures  are  much  less.  It  would  also  be  reasonable 
to  assume  less  ice  on  the  contact  wire  than  on  steel  carrying  strands, 
especially  on  lines  of  large  traffic.  Mr.  Coombs'  recommendations, 
in  regard  to  wind  pressures  and  unit  strains,  are  generally  reasonable. 
His  paper,  however,  in  common  with  most  writings  on  the  same  sub- 
ject, suffers  from  an  insufficient  consideration  of  the  bending  strains 
in  the  wires.  In  many  designs,  these  bending  strains  are  greater  than 
the  tensions,  and  their  neglect  leads  inevitably  to  the  selection  of  un- 
safe designs.  They  vary  so  greatly  in  amount  that  they  cannot  be 
provided  for  by  neglecting  them  and  adopting  a  large,  but  uniform 
factor  of  safety. 

*This  discussion  (of  the  paper  by  R.  D.  Coombs,  M.  Am.  Soc.  C.  E.,  printed  in  Pro- 
ceedings for  December,  1007),  is  printed  in  Proceedings  in  order  that  the  views  expressed 
may  be  brought  before  all  members  for  further  discussion. 


230  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Mayer.  Mr.  Coombs  suggests  for  discussion  a  suspension  from  a  single 
steel  strand  with  300-ft.  spans  and  a  distance  of  6J  ft.  from  the  points 
of  suspension  of  the  strand  to  the  wire.  Taking  the  wire  as  horizontal, 
and  at  the  highest  temperature,  this,  with  a  distance  of  6  in.  from 
the  lowest  point  of  the  strand  to  the  wire,  gives  a  maximum  vertical 
deflection  of  6  ft.  for  the  strand.  He  prefers  this  to  the  double 
catenary  and  strandless  suspension.  What  justifies  this  preference, 
and  what  makes  a  superior  suspension? 

To  have  adequate  conductivity,  the  wire  must  be  copper;  to  give 
a  smooth  path  for  the  bow,  it  must  be  solid;  it  is  impossible  to 
make  it  straight,  but,  under  the  pressure  of  the  sliding  bow,  especially 
with  high  train  speeds,  it  must  have  a  large  minimum  radius  of  verti- 
cal curvature.  The  wire  should  not  deviate  horizontally  far  from  the 
center  line  of  track,  or,  if  hung  on  the  side,  from  a  line  parallel  to  it. 
It  must  remain  safely  suspended  under  the  influence  of  its  weight  and 
that  of  sleet,  the  wind  pressure,  the  prjessure  of  the  sliding  bow,  and 
the  changes  of  temperature.  The  wire  and  its  supporting  structure 
should  interfere  as  little  as  possible  with  the  view  of  the  signals. 
These  ends  should  be  attained  by  the  simplest  means,  entailing  the 
least  cost  of  construction  and  maintenance.  The  bow  lifts  the  wire, 
and  the  curvature  of  its  motion,  and  not  that  of  the  freely  hanging 
wire,  must  be  considered.  If  the  wire  is  supported  at  short  intervals, 
it  is  lifted  by  the  passing  bow  to  positions  above  its  supports;  if  these 
are  rigid,  the  bow,  at  high  train  speed,  oscillates  rapidly  up  and  down. 
Excessive  bending  strain  in  the  wire,  and  jumping  of  the  bow,  with 
sparking,  may  result. 

To  ascertain  whether  the  bow  will  run  smoothly,  its  equivalent 
weight,  the  train  speed,  the  cross-section  of  the  contact  wire  and  its 
tension,  the  distance  apart  of  the  suspenders,  their  weight,  and  that 
of  the  carrying  strand,  and  the  nature  of  the  connection  of  the  wire 
to  the  suspenders  must  be  known.  For  high  train  speeds,  a  small 
equivalent  weight  of  the  bow  is  essential.  Long  bows  are  inevitably 
heavy;  high  train  speeds,  therefore,  require  short  bows  and  small 
lateral  deflections  of  the  contact  wire.  The  equivalent  weight  of  the 
bow,  however,  depends  even  more  on  the  design  adopted  than  on  its 
length.  If  smooth  running  is  called  for,  at  high  speed,  with  an  in- 
ferior heavy  bow,  short  spans  are  inevitable  with  all  suspensions.  The 
rapid  vertical  oscillation  of  the  bow  is  avoided  in  the  strandless  and 
the  Siemens-Schuckert  suspension  described  by  Mr.  Coombs. 

In  the  strandless  suspension,  with  long  spans,  there  is  a  large 
change  in  the  direction  of  motion  of  the  bow  at  the  infrequent  sus- 
penders; to  make  this  practicable  at  high 'speed,  the  curvature  of  this 
motion  must  be  chosen  so  that  where  it  is  convex  downward  the  bow 
will  not  jump,  and  where  it  is  convex  upward  neither  the  wire  nor  the 
bow  nor  the  suspender  will  suffer  from  the  increased  pressure.    In  all 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTKIC  TRACTION  231 

suspensions  there  are  large  changes  in  the  direction  of  motion  of  Mr.  Mayer, 
the  bow  at  low  overhead  crossings  and  tnnnel  entrances,  and  some- 
times at  grade  crossings.  With  inferior,  heavy  sliding  bows  a  perfect 
design  of  the  contact  line  for  high  speeds  at  these  points  is  difficult 
or  impracticable.  A  suspender  fitted  to  ch.^nge  the  variable  direction 
of  approach  of  the  sliding  bow  into  another  variable  direction  of  its 
departure,  by  a  transition  curve  of  large  least  radius  at  all  tempera- 
tures, is  here  needed  with  all  suspensions. 

To  obtain  a  safe  wire,  its  maximum  strain  must  nowhere  and 
never  exceed  about  three-fourths  of  its  elastic  limit.  It  is  exposed 
to  bending  strains  and  tensions.  The  former  are  often  much  larger 
than  the  latter.  In  catenary  suspensions,  a  large  grooved  copper  wire, 
about  0.3  in.  wide  and  0.6  in,  in  height  of  cross-section,  is  suspended 
from  steel  strands  made  up  of  very  small  wires.  The  copper  wire  has 
an  ultimate  strength  of  from  50  000  to  60  000  lb.,  the  steel  wires, 
140  000  lb.  or  more.  The  modulus,  E,  of  copper  wires  is  16  000  000, 
that  of  steel  strands  is  26  000  000  lb.  per  sq.  in.  of  solid  section.  It 
is  evident  that  the  bending  accompanying  changes  in  vertical  and 
horizontal  deflections  will  produce  much  more  serious  bending  strains 
in  the  large  and  weak  copper  wire  than  in  the  small  and  strong  steel 
wires.  For  calculating  them,  the  vertical  and  horizontal  deflections 
of  the  wire  under  all  conditions  of  load,  wind  pressure,  pressure  of  the 
sliding  bow,  and  changes  of  temperature  must  be  determined.  It  is 
easy  to  provide  steel  ropes  strong  enough  to  carry  the  wire,  the  ice 
loads  and  the  wind  pressures.  The  main  difficulty  arises  from  their 
expansion  and  contraction  caused  by  changes  of  temperature  and  ten- 
sion. These  and  the  wind  pressures  cause  lateral  and  vertical  curva- 
ture of  the  contact  wire.  Both  the  deflections  and  the  drop  of  tem- 
perature increase  its  tension  and  produce  at  certain  points  large  bend- 
ing strains.  To  determine  the  degree  of  safety  of  the  various  sus- 
pensions, the  largest  bending  strains  and  tensions  in  the  contact  wire 
and  the  ropes  must  l)e  calculated.  With  regard  to  obstruction  to  the 
view  of  the  signals,  1h(^  fewer  ropes,  suspenders,  posts,  and  bridges  or 
brackets,  the  better. 

For  judging  the  suspension  suggested  by  Mr.  Coombs  by  these 
standards,  and  for  finding  whether  it  is  sufficiently  rigid  to  be  suit- 
able for  use  with  a  sliding  bow  that  will  run  smoothly  at  high  speed, 
and  sufficiently  strong  to  resist  with  adequate  safety  the  incident 
forces,  the  size  of  the  steel  strand  and  the  distance  and  nature  of  the 
suspenders  must  be  reasonably  assumed,  and  the  deflections  and  con- 
sequent bending  strains  and  tensions  calculated. 

Following  Mr.  Coombs'  specification,  a  |--in.  steel  strand  is  ample, 
and  weighs  0.89  lb.  per  ft.  of  span.  The  0000  copper  wire  weighs 
0.64  lb.,  and  gas-pipe  suspenders,  12  ft,  apart,  about  0.33  lb.,  giving 
a  total  weight  of  1.86  lb.  per  ft.  of  span.    Ice  ^  in.  thick  on  all  parts 


232  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Mayer,  weighs  1.63  lb.,  giving  a  total  weight,  with  ice,  of  3.49  lb.  per  ft.  of 
span.  Taking  the  wire  to  consist  approximately  of  two  0  wires 
directly  above  each  other,  it  is  0.65  in.  high  and  half  as  wide.  The 
wind  pressure  on  the  bare  metal  of  wire  suspenders  and  strand,  taking 
12  lb.  per  sq.  ft.  all  through,  is  1.63  lb.,  and  that  on  the  ice-covered 
structure,  with  8  lb.  per  sq.  ft.,  2.57  lb.  per  ft.  of  span.  Assuming  a 
tension  of  1  000  lb.  in  the  contact  wire,  the  lateral  deflection  of  the 
wire,  at  maximum  temperature,  woi;ld  be  3.15  ft.  This  assumes  that 
the  wire  is  held  at  the  ends  of  the  spans  by  steady  braces.  If  these 
were  absent,  the  lateral  deflection  would  be  much  larger  and  thq 
needed  sliding  bow  would  be  altogether  impracticable. 

Where  sliding  bows  are  used,  the  wire  must  run  alternately  to  the 
right  and  left  of  the  center  line  of  track  so  as  to  distribute  the  wear 
over  a  considerable  length  of  the  bow.  Taking  the  lateral  displace- 
ment of  the  wire  at  the  brackets  to  be  1  ft.,  and  allowing  i  ft.  for  the 
lateral  vibration  of  the  sliding  bow,  the  latter  must  be  6.96  ft.  long 
to  catch  the  wire,  with  the  strongest  winds  assumed.  A  sliding  bow 
of  this  length,  which  will  not  jump  at  the  suspenders,  with  moderate 
train  speeds,  can  be  designed. 

For  the  highest  present  steam  railway  speeds,  a  much  larger  ten- 
sion in  the  contact  wire  or  more  frequent  suspenders  are  needed  to 
prevent  jumping  and  sparking,  with  the  usual  connection  of  the  wire 
to  the  suspender.  Jumping  and  sparking  might  also  be  prevented  by 
a  contrivance  allowing  the  wire  to  rise  at  the  suspenders,  without 
lifting  them,  when  the  sliding  bow  passes.  With  steady  braces,  which 
are  practically  unavoidable  with  this  design,  the  wire  carries,  at 
maximum  temperature,  0.49  lb.  of  the  total  wind  pressure  of  1.63  lb. 
per  ft.,  to  the  steady  braces,  and  its  tension  is  thereby  increased. 
Much  more  serious  is  the  bending  strain  in  the  wire  at  the  clamps 
which  connect  it  to  the  steady  braces.  These  clamps  may  be  designed 
to  avoid  bending  strain  in  the  wire  at  maximum  temperature  without 
wind.  In  this  case,  the  lateral  bending  strain  in  the  wire  at  the  end 
of  the  clamp,  at  highest  temperature  and  wind  pressure,  is  35  300  lb. 
per  sq.  in.     For  the  wire  here  assumed,  this  bending  strain  is  given 

19  200  0  7i       ,  •     ^1      1      J-  .     •       • 

by  the  formula,  s  =    7- where  s  is  the   bending  strain,  m 

V  T 
pounds  per  square  inch,  and   T  is  the  tension   in  the  wire.     If  T  is 
decomposed  into  a  component  having  the  direction  of  the  wire  at  the 
end  of  the  clamp  and  one  normal  to  it,  the  horizontal  component  of 
the  latter  is  Q  h,  and  the  vertical  component  Q  v.     For  the  vertical 

18  000  Q  V 

bending  of  the  same  wire,  the  formula  is  s  =    /-  At    the 

V  T 

same  time,  with  this  bending  strain  of  35  300  lb.  per  sq.  in.,  the  tension 

in  the  wire  is  10  800  lb.,  giving  a  combined  strain  of  46  100  lb.  per 

sq.  in. 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  333 

The  passing  sliding  bow  increases  the  total  strain  to  nearly  48  000  Mr.  Mayer, 
lb.  per  sq.  in.  The  horizontal  bending  strain  may  be  rediiced,  theoreti- 
cally, to  one-half,  by  using  two  steady  braces  at  each  bracket  and  con- 
necting them  to  the  wire  by  hinged  clamps,  the  hinges  being  vertical. 
The  steady  braces  themselves  should  have  horizontal  hinges  permitting 
the  clamps  to  rise  and  fall  with  changes  of  temperature  and  wind 
pressure  and  the  passing  of  the  sliding  bow.  By  this  rise  and  fall,  the 
principal  vertical  bending  strains  of  the  wire,  except  those  due  to  the 
passing  bow,  are  transferred  to  the  nearest  suspender;  therefore,  they 
need  not  be  added  to  the  horizontal  bending  strains  occurring  at  the 
steady  braces.  With  hinged  double  steady  braces,  connected  to  the 
wire  by  hinged  clamps,  the  total  strain  per  square  inch  at  the  highest 
temperature,  with  the  assumed  wind  pressure,  is  approximately  30  500 
lb.  This  assumes  that  all  the  hinges  work  without  friction.  The  fric- 
tion of  the  hinges  may  increase  this  strain  considerably.  At  the 
lowest  temperature,  without  ice,  and  with  the  largest  wind  pressure, 
the  lateral  deflection  of  the  wire  is  15  ft.,  its  tension  is  5  000  lb.,  or 
30  080  lb.  per  sq.  in. 

The  bending  strain  at  the  steady  braces  due  to  horizontal  bending 
is  27  200  lb.  per  sq.  in.  with  single,  and  half  as  much  with  double, 
braces  and  hinged  clamps.  This  gives  combined  strains  of  57  280  and 
43  680  lb.  per  sq.  in.  These  strains  are  both  increased  about  1  000  lb. 
by  the  passing  sliding  bow.  Since  they  exceed  the  elastic  limit,  the 
wire  will  bend  before  the  strains  reach  the  amount  calculated.  Re- 
peated forward  and  backward  bending  will  produce  rupture.  To  re- 
duce these  large  bending  strains  and  tensions  in  the  contact  wire, 
smaller  deformations  must  be  obtained.  These  can  be  secured  by 
smaller  deflections  of  the  carrying  strand  which  requires  either  heavier 
strands  or  shorter  spans. 

With  a  |-in.  strand  of  3  ft.  maximum  vertical  deflection  and  300 
ft.  span,  carrying  6  in.  below  the  strand  at  the  center  of  the  span  a 
0000  grooved  wire  which  is  horizontal  and  has  1 000  lb.  tension 
at  the  highest  temperature,  without  wind,  the  deflections  and  strains 
in  the  wire  are  as  follows:  At  the  highest  temperature  and  wind 
pressure,  the  horizontal  deflection  of  the  wire  is  2.17  ft.,  its  upward 
deflection  is  0.23  ft.,  the  tension  is  8  260  lb.  per  sq.  in.,  the  horizontal 
bending  strain  is  20  580  lb.  per  sq.  in.  with  single  and  half  as  much 
with  double  steady  braces.  This  gives,  with  the  latter,  a  combined 
strain  of  18  550  lb.,  which  is  increased  to  about  20  600  lb.  per  sq.  in. 
by  the  passing  sliding  bow.  This  is  a  great  improvement  over  the 
corresponding  30  500  lb.  with  6  ft.  deflection  of  the  strand. 

At  the  lowest  temperature,  with  the  highest  wind  pressure,  without 
ice,  the  vertical  deflection  of  the  strand  is  1.85  ft.,  its  lateral  deflec- 
tion is  0.9  ft.  The  vertical  upward  deflection  of  the  wire  is  1.18  ft. 
and  its  lateral  deflection  1.05  ft.     The  tension  in  the  wire  is  4  760  lb.. 


234  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Mayer,  or  28  640  lb.  per  sq.  in. ;  the  horizontal  bending  strain,  with  double 
steady  braces,  is  9  290  lb.,  giving  a  combined  strain  of  37  930  lb.  per 
sq.  in.  This  is  increased  to  about  39  000  lb.  by  the  passing  sliding 
bow.  With  hinged  steady  braces,  most  of  the  vertical  bending  strain, 
amounting,  if  concentrated,  to  24  800  lb.  per  sq.  in.,  is  transferred  to 
several  of  the  nearby  suspenders.  If  these  suspenders  have  clamps  at- 
tached to  them  by  horizontal  hinges,  allowing  oscillation  of  the  wire 
in  a  vertical  plane,  then  the  vertical  bending  strains  are  certainly 
smaller  than  the  horizontal  ones;  if  there  are  no  such  hinges,  such 
strains  are  probably  larger  than  the  horizontal  bending  strains  with 
double  steady  braces,  but  they  cannot  easily  be  calculated  with  ac- 
curacy. With  ice  and  wind,  at  the  lowest  temperature,  the  lateral 
deflection  of  the  wire  is  1.48  ft.,  the  upward  deflection  is  0.46  ft.,  its 
tension  is  4  760  lb.  or  28  640  lb.  per  sq.  in.  The  horizontal  bending 
strain,  with  double  steady  braces  and  hinged  clamps,  is  13 100  lb., 
giving  a  combined  strain  of  41  740  lb.  per  sq.  in.  This  is  increased 
about  1  000  lb.  by  the  passing  bow. 

The  maximum  tension  in  the  strand  is  16  200  lb.,  which  gives  a 
factor  of  safety  of  3,  provided  the  small  bending  strains  are  neglected. 
A  sliding  bow  5  ft.  long  is  needed,  and  can  be  designed  so  as  to  give 
smooth  running  at  all  but  the  highest  speeds.  It  is  evident  that  the 
strains  in  the  wire  are  still  excessive  where  the  wind  pressures  and 
ice  loads  prescribed  by  the  specification  really  occur.  As  these  are  of 
rare  occurrence,  a  structure  of  this  design,  with  improved  hinged 
double  steady  braces,  connected  to  the  wire  by  hinged  clamps,  will 
probably,  in  most  situations,  last  a  number  of  years.  It  would  have 
about  the  rigidity  of  a  double  catenary  suspension  of  the  same  span 
with  two  strands  of  A  in.  diameter  and  6  ft.  vertical  deflection, 
having  the  wire  6  in.  below  the  lowest  point  of  the  strands.  The 
largest  lateral  deflection  of  the  wire  of  this  latter  suspension,  with  the 
wire  tension  and  wind  pressure  here  assumed,  is  approximately  2.25 
ft.  In  the  double  catenary  suspension,  no  steady  braces  are  used,  the 
bending  strains  in  the  wire  due  to  its  vertical  and  lateral  deflection 
are  distributed  to  several  clamps  near  the  ends  of  the  spans.  They 
cannot  easily  be  calculated  with  accuracy,  but  are  probably  somewhat 
smaller  than  in  the  best  single  catenary  suspension  of  the  same  span. 
The  tensions  in  the  wire  are  nearly  the  same  in  both  designs  here 
compared.  The  double  catenary  suspension  is  certainly  superior  in 
strength  to  a  single  catenary  suspension  of  the  same  span  and  lateral 
deflection  with  single  steady  braces.  All  these  designs  are  far  in- 
ferior in  safety  to  railroad  bridges. 

Taking  now  a  0000  round  wire,  hung  from  special  suspenders, 
with  300-ft.  spans  and  4  ft.  maximum  vertical  deflection,  with  strain 
adjusters  1  mile  apart,  the  adjusters  changing  the  length  of  the 
spans  four  times  a  year,  so  that  the  variation  of  temperature  with  one 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  235 

length  of  span  does  not  exceed  84°  f ahr. :  The  largest  lateral  deflec-  Mr.  Mayer, 
tion  of  the  wire  is  2.54  ft.  A  sliding  bow  6  ft.  long  is  required.  The 
tension  in  the  wire  at  the  lowest  temperature,  with  wind  and  without 
ice,  is  3  260  lb.  or  19  600  lb.  per  sq.  in.  The  maximum  bending  strain 
at  the  same  time  is  6  250  lb.,  giving  a  total  strain  of  25  850  lb.  per 
sq.  in.  The  corresponding  strain  in  the  suspension  suggested  by  Mr. 
Coombs  is  44  680  lb.  per  sq.  in.  In  the  single  catenary  suspension, 
with  3  ft.  maximum  vertical  deflection  of  the  strand,  it  is  39  000  lb. 
with  the  best  design. 

With  a  coating  of  ice,  i  in.  thick  on  the  contact  wire,  increasing  its 
diameter  i  in.,  the  maximum  tension  at  the  lowest  temperature,  and 
with  a  wind  pressure  of  8  lb.  per  sq.  ft.  of  ice-covered  wire,  is  21  530 
lb.,  the  bending  strain  is  6  250  lb.,  giving  a  total  of  27  780  lb.  per  sq. 
in.  A  greater  thickness  of  ice  on  the  contact  wire  would  make  it 
difBcult  to  collect  the  current.  Where  there  is  considerable  traffic, 
the  wire  will  be  generally  several  degrees  warmer  than  the  atmosphere, 
and  less  ice  will  form  on  it  than  on  steel  strands  carrying  but  little 
current,  and  the  passing  sliding  bow  will  kneck  off  much  of  that 
which  forms.  It  is  reasonable,  therefore,  to  assume  a  smaller  amount 
of  ice  on  the  contact  wire  than  on  the  strands.  The  maximum  strain 
in  the  wire,  with  ice  i  in.  thick,  and  a  wind  pressure  of  8  lb.  per  sq. 
ft.,  at  the  lowest  temperature,  in  the  best  of  the  single  catenary  suspen- 
sions of  the  same  span  is  42  700  lb.  per  sq.  in. ;  this  would  be  but 
little  reduced  by  assuming  the  ice  on  the  contact  wire  i  in.  thick,  the 
strain  without  any  ice  being  39  000  lb. 

The  strandless  suspension  here  described  requires,  for  smooth  run- 
ning with  a  speed  of  70  miles  per  hour,  a  sliding  bow  of  4  lb.  equiva- 
lent weight,  6  ft.  long.  Such  a  bow  can  easily  be  designed,  but,  as 
far  as  the  speaker  is  aware,  it  is  not  at  present  in  the  American  mar- 
ket. The  bows  in  use  are  designed  for  smaller  speeds.  If  they  are  to 
be  used  with  high  speeds,  shorter  spans  are  necessary.  The  4-ft.  de- 
flection of  the  contact  wire  requires  a  larger  range  of  vertical  motion 
of  the  bow  than  the  catenary  suspension  of  the  same  span,  in  which 
the  height  of  the  wire  varies  only  about  2  ft.  Though  300-ft.  spans 
are  entirely  practicable  and  safe,  with  improved  strandless  suspension 
and  a  speed  of  70  miles  per  hour,  they  will  riot  give  continuous  con- 
tact at  this  speed  without  improved  sliding  bows. 

The  speaker  has  invented  another  suspender,  which  can  be  used 
at  any  speed  with  ordinary  sliding  bows  of  large  equivalent  weight, 
and  reduces  still  further  the  bending  strains.  With  it,  300-ft.  spans 
can  be  safely  used.  As  the  sliding  bow  may  be  heavy,  it  may  be  made 
longer,  and  a  maximum  vertitial  deflection  of  4J  or  5  ft.  may  be 
adopted,  thus  reducing  greatly  its  maximum  tension.  This  suspender 
will  be  described  later. 

The    calculation    showing    the    excessive    bending    strains    in    the 


236  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Mayer,  single  catenary  sv;spensions  of  300-ft.  span  is  confirmed  by  practical 
experience  with  long-span  electric  transmission  lines.  In  these,  solid 
wires  were  first  used,  but  they  broke  at  the  insulators  even  with 
moderate  tension  per  square  inch.  Stranded  M'ires,  therefore,  are 
now  used  with  long  spans.  Many  experiments  have  been  made  with 
single  catenary  suspensions.  As  a  result,  the  present  practice  in 
America  and  in  Europe,  as  far  as  known  to  the  speaker,  does  not 
show  any  existing  spans  of  more  than  160  ft.  The  importance  of  re- 
ducing the  lateral  and  vertical  deflection  of  the  contact  wire,  and 
thereby  its  bending  strains,  is  fully  appreciated  by  the  designers  of 
most,  if  not  all,  of  the  existing  structures  in  America.  The  sliding 
bows  are  generally  4  ft.  long,  and  could  not  be  used  with  large  lateral 
deflections. 

Taking  a  design  with  150-ft.  spans  of  a  steel  strand  of  ^^  in.  diam- 
eter, with  16  in.  maximum  vertical  deflection,  the  0000  grooved 
wire  being  horizontal  at  the  highest  temperature  and  4  in.  below  the 
strand  at  the  center  of  the  span,  the  following  deflections  and  strains 
are  obtained  with  a  variation  of  temperature  of  140°  fahr.,  and  the 
loads  and  wind  pressures  mentioned  by  Mr.  Coombs :  The  weight  of 
wire,  strand  and  suspenders  is  approximately  1.04  lb.  per  ft.,  the  wind 
pressure,  without  ice,  is  1.15  lb.  per  ft.  The  weight,  with  ice,  is  2.33 
lb.,  and  the  wind  pressure  is  2.10  lb.  per  ft.  With  the  greatest  wind 
pressure,  and  at  the  highest  temperature,  the  lateral  deflection  of  the 
wire  is  approximately  0.95  ft.,  its  tension  is  1  284  lb.,  or  7  730  lb.  per 
sq.  in.,  the  horizontal  bending  strain,  with  ordinary  steady  braces,  is 
17  410  lb.,  giving  a  combined  strain  of  25  140  lb.  per  sq.  in. ;  this  is 
increased  to  about  27 100  lb.  by  the  passing  of  an  improved  sliding 
bow  of  a  maximum  dynamic  pressure  of  25  lb. 

At  the  lowest  temperature,  with  the  greatest  wind  pressure,  and 
without  ice,  the  upward  deflection  of  the  wire  is  approximately  0.3  ft., 
the  horizontal  deflection  is  0.38  ft.,  its  tension  is  4  640  lb.,  or  27  920 
lb.  per  sq.  in.  The  horizontal  bending  strain  is  6  630  lb.,  giving  a 
combined  strain  of  34  550  lb.  per  sq.  in.  This  is  increased  to  about 
35  500  lb.  by  the  passing  sliding  bow.  If  the  wire  is  firmly  held  at 
the  steady  brace,  so  that  it  cannot  rise  and  fall,  the  combined  strain 
due  to  tension  and  horizontal  and  vertical  bending  is  about  40  000  lb. 

With  ice  having  an  average  thickness  of  i  in.,  at  lowest  tempera- 
ture and  with  the  greatest  wind  pressure,  the  wire  has  a  lateral  deflec- 
tion of  0.6  ft.  and  a  downward  deflection  of  0.07  ft.  Its  tension  is  28  160 
lb.,  and  the  horizontal  bending  strain  is  10  510  lb.,  giving  a  combined 
strain  of  38  670  lb.  per  sq.  in.  This  is  increased  by  about  1 000  lb. 
per  sq.  in.  by  the  passing  sliding  bow.*  The  horizontal  bending  strain 
may  be  reduced  to  one-half  and  the  vertical  bending  strain  transferred, 
by  a  perfect  double  steady  brace,  allowing  the  wire  to  rise  and  fall 
and  turn. 


Papors.]  DISCUSSION  ON  OVERHEAD  ELECTRIC   TRACTION  237 

Wire  havino-  an   elastic  limit  of  40  000  lb.  per  sq.  in.  can  be  ob-  Mr.  Mayer, 
tained,  and.  with  little  ice  and  moderate  wind  pressures  and  changes 
of  temperature,  jjives  an  approximately  safe  structure  with  the  usual 
designs. 

The  maximum  tension  in  the  steel  strand,  with  the  foregoing 
loads,  is  5  000  lb.,  the  ultimate  strength  of  a  cast-steel  strand  is  about 
13  600  lb. 

These  results  explain  why  much  longer  spans  are  not  used  with 
single  catenary  suspension. 

It  is  evident  that  a  structure  having  the  factor  of  safety  of  a 
railroad  bridge  is  not  practicable  with  ordinary  catenary  suspension, 
in  most  climates.  If  Mr.  -Coombs'  description  of  the  Siemens-Schuc- 
kert  suspension  is  correct,  and  if  the  pulleys  over  which  the  con- 
tact wire  is  carried  have  a  diameter  of  12  ft.,  larger  pulleys  being 
impracticable,  the  bending  strain  in  the  wire  here  assumed  would  be 
36  000  lb.  per  sq.  in.  If  the  tension  in  the  contact  wire  is  made  small, 
the  lateral  deflection  of  the  wire  would  be  much  increased,  and  long 
spans  with  large  strand  deflection  would  be  impracticable.  This  sus- 
pension is  used  with  spans  of  48  m.,  with  a  steel  carrying  structure 
where  long  spans  are  very  desirable.  If  the  designers  had  thought 
them  practicable,  they  would  probably  have  adopted  them.  Spans  of 
300  ft.,  with  single  catenary  suspension,  therefore,  are  not  sustained 
by  precedent;  they  cannot  be  defended  successfully  by  theory,  and 
they  will  probably  prove  short-lived  if  tried  under  conditions  ap- 
proximating those  here  assumed. 

The  maximum  strains  which  demonstrably  exist  in  the  contact 
wires  with  catenary  suspension  show  what  a  copper  wire  can  stand,  at 
least  for  a  few  years.  They  make  it  extremely  probable  that  a  wire  in 
which  the  maximum  strain  never  and  nowhere  exceeds  30  000  lb.  per 
sq.  in.  is  abundantly  safe. 

W.  K.  Archbold,  Esq. — Regarding  the  matter  of  protective  struc-  Mr.  Archbold. 
tures  where  transmission  lines  cross  railroad  tracks,  Mr.  Coomb's 
paper  should  help  to  standardize  the  practice,  which  has  varied  ex- 
tremely, as  the  speaker  has  had  occasion  to  note.  Under  the  direction 
of  Thomas  H.  Mather,  M.  Am.  Soc.  C.  E.,  an  overhead  construction 
has  recently  been  designed  and  installed  on  the  line  of  the  Syracuse, 
Lake  Shore  and  Northern  Railroad,  running  from  Syracuse  to 
Baldwinsville,  N.  Y.  The  line  is  about  5  miles  long,  and  is  provided 
with  single-catenary  trolley  construction  presenting  some  new  features, 

Mr.  Mather  thinks  that  the  work  has  not  yet  advanced  far  enough 
to  warrant  the  presentation  of  a  formal  paper,  and  therefore  the 
speaker  will  make  simply  a  preliminary  presentation  of  the  prominent 
features  of  the  construction. 

The  trolley  wire  is  hung  from  a  messenger  cable  supported  on 
bridges  spaced  300  ft.  apart,  from  center  to  center.     The  bridges,  as 


238  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Archboid.  shown  ill  Plate  XXXIV,  consist  of  light  trusses  on  bents  30  ft.  apart, 
from  center  to  center.  The  bents  are  each  composed  of  two  8-in.  chan- 
nels, 6  ft.  apart  at  the  base,  converging  to  8  in.  at  the  top,  and  sup- 
ported on  concrete  pedestals,  20  in.  square,  and  of  depth  varying  with 
the  nature  of  the  ground.  The  trusses  have  an  8-in.  channel  top  chord 
and  6-in.  channel  bottom  chords,  set  with  the  flanges  down.  The 
diagonal  members  are  g-in.  rods,  and  the  struts  2J  by  2J  by  i-in. 
angles.  The  struts  are  flattened  and  bent  over  at  the  ends,  and  are 
riveted  to  the  channels. 

To  the  top  chord  of  each  truss  are  bolted  malleable-iron  pins  to 
v/hich  are  cemented  porcelain  insulators  for  the  messenger  cable.  The 
three-phase  high-tension  line  is  supported  on  steel  A-frames  at  each 
end  of  each  bridge.  The  construction  is  designed  for  a  wind  pressure 
of  8  lb.  per  sq.  ft.  on  the  trolley  and  messenger  cables,  covered  with 
i  in.  of  ice,  a  somewhat  lower  ice  load  being  assumed  on  the  high- 
tension  cables,  which  are  of  No.  2  copper.  The  structure  is  computed 
as  a  braced  portal,  the  unit  strains  under  the  assumed  wind  and  ice 
load  being  22  500  lb.  per  sq.  in.,  reduced  for  compression  members. 

The  catenary  is  strung  for  a  net  sag  of  6.5  ft.  at  100°  fahr.  At  20° 
fahr.,  the  sag  is  about  6.5  ft.,  and  the  trolley  is  about  1  ft.  higher  at 
the  center  of  the  span  than  under  the  bridges,  the  height  from  rail  to 
trolley  being  18  ft.  at  the  bridges.  Stranded  steel  messenger,  15  000-lb. 
wire,  y^  in.  in  diameter,  supports  steel  hangers,  |  in.  in  diameter, 
spaced  10  ft.  apart,  from  center  to  center.  These  hangers  are  of  the 
Ohio  Brass  Company  type,  and  are  attached  to  the  messenger  cable 
with  a  sister  hook  through  the  base  of  which  the  rod  is  threaded  and 
drawn  up  tight  against  the  messenger  cable.  The  0000  grooved 
trolley  is  secured  to  the  hanger  by  Detroit  clamps. 

At  each  bridge  there  is  a  span-wire  steady  strain  (not  shown  in  the 
photograph),  the  trolleys  being  insulated  from  the  bridge  and  from 
each  other  by  6-in.  wheel-type  porcelain  strain  insulators.  The  messen- 
ger cable  is  dead-ended  on  an  equalizer  attached  to  a  pair  of  these  in- 
sulators, which  are  connected  by  short  cable  loops  to  a  similar  pair 
secured  to  the  anchor  bridge. 

A  test  which  came  on  the  line  during  construction  showed  the 
effect  of  a  broken  messenger  wire.  When  about  half  a  mile  of  wire 
had  been  pulled  up,  and  hangers  were  on  three  or  four  of  the  spans, 
the  dead-end  arrangement  broke,  allowing  the  line  to  go.  The  insu- 
*  lator  broke  on  the  bridge  next  to  the  dead-end  bridge,  but  the  trouble 

did  not  extend  any  further  than  that  point.  Some  of  the  men  on  the 
work  thought  the  foundation  of  the  second  bridge  was  raised  a  little, 
but  there  seems  to  be  considerable  difference  of  opinion  on  that  point, 
and  certainly  no  damage  was  done.  The  idea  has  been  that  the  bridges 
would  be  what  might  be  called  semi-anchored;  or,  in  other  words,  the 
effect  of  a  break  would  not  go  beyond  two  or  three  bridges  in  either 


PLATE  XXXIV. 

PAPERS,   AM.  SOC.  C.   E 

MARCH,   1908. 

ARCHBOLD    ON 

OVERHEAD  CONSTRUCTION 

FOR    ELECTRIC    TRACTION. 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  239 

direction.     It  is  possible,  too,  that  the  heavy  rods  and  tight  clamping  Mr.  Archboid. 
effect,  obtained  both  at  the  messenger  and  trolley  wires,  assisted  con- 
siderably in  holding  the  line.    At  any  rate,  no  damage  was  done  which 
could  not  be  repaired  quickly. 

The  high-tension  wire  is  strung  with  a  sag  of  40  in.  at  20°  fahr., 
and  has  a  net  clearance  of  about  24  ft.  from  the  track.  As  the  trolley 
may  at  some  time  be  operated  at  6  600  volts,  single-phase,  all  the  in- 
sulation of  the  catenary,  steady  strains,  etc.,  is  designed  to  withstand 
this  voltage.  At  present,  however,  it  is  being  operated  at  600  volts, 
direct-current.  In  regard  to  the  lateral  stiffness  of  the  single  catenary 
with  supports  300  ft.  apart,  it  may  be  noted  that  the  line  has  been  in 
operation  about  10  days,  and,  thus  far,  very  little  deflection  or  rolling- 
can  be  observed  under  the  action  of  the  wheel  trolley  with  a  tension  of 
about  25  lb.  During  this  time  there  have  been  temperature  changes 
of  50°  and  wind  velocities  of  45  miles  per  hour.  These  conditions 
seemed  to  make  no  difference  in  the  operation,  even  before  the  steady 
strains  were  installed. 

The  speaker  feels  justified  in  saying  positively  that,  with  this  type 
of  construction,  there  will  be  no  difficulty  from  side-sway  on  the  300-ft. 
span.  It  is  yet  to  be  determined  whether  the  line  is  too  stiff  in  the 
vertical  plane,  but  that  cannot  be  determined  positively  except  by 
operation  extending  over  a  considerable  period,  and  including  hot  as 
well  as  cold  weather.  The  preliminary  tests  which  have  been  made  in- 
dicate that  there  will  be  no  difficulty. 

This  line  is  a  re-location,  to  shorten  the  running  time  and  provide 
a  double  track  on  private  right  of  way  between  Syracuse  and  Baldwins- 
ville.  The  old  single-track  line  is  on  a  highway,  and  is  about  |  mile 
longer  than  the  new  location,  which  will  form  part  of  a  new  high-speed 
electric  road  between  Syracuse,  Fulton,  and  Oswego,  a  total  distance 
of  about  35  miles. 

Charles  Eufus  Harte,  M.  Am.  Soc.  C.  E. — In  the  field  of  trans-  Mr.  Harte. 
mission-  and  distribution-line  construction,  each  engineer  lias  been 
largely  a  law  unto  himself,  and  Mr.  Coombs'  effort  to  secure  some 
measure  of  standardization  is  much  to  be  commended.  At  the  same 
time,  local  conditions  very  largely  govern,  and  the  successful  construc- 
tion of  one  locality  may  be  of  little  value  even  in  comparatively  near 
sections. 

In  addition  to  the  causes  given  by  the  author,  a  short  circuit  may 
be  caused  by  the  swaying  together  of  two  phases  of  the  circuit.  This, 
however,  may  be  prevented  by  spacing  the  wires  a  distance  apart  equal 
to  at  least  twice  the  versed  sine  of  the  sag.  Thus  the  Missouri  River 
transmission  has  a  spacing  of  78  in. ;  the  1  450-f t.  Connecticut  River 
span  of  the  Springfield-Suffield  Line,  84  in.;  while  the  Madison  River 
Line,  of  Montana,  has  108  in.  As  an  additional  precaution,  the  wires 
are  often  arranged  so  that  no  two  are  in  the  same  horizontal  plane. 


240  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Harte.  This  is  characteristic  of  the  Connecticut  Eiver  span,  the  Anglo- 
Mexican,  the  Southern  Power,  and  many  other  transmission  lines. 
The  large  spacing  (of  the  triangle),  with  two  wires  in  a  vertical  plane, 
also  materially  reduces  the  likelihood  of  interference  from  large  birds, 
or  branches  or  wires  blown  against  the  line.  In  practice,  however,  be- 
cause of  their  weight  and  low  periodicity,  long  spans  usually  swing 
in  unison,  thus  maintaining  the  spacings. 

Mr.  Coombs  states  that  insulator  troubles  are  largely  due  to  mis- 
directed savings.  While  it  is  true  that  there  are  to-day  many  lines 
operating  without  change  at  higher  voltages  than  designed  for,  and 
on  which  the  insulators  were  poor  for  the  original  voltage,  there  are 
many  other  lines  where,  although  no  expense  has  been  spared,  insulator 
troubles  are  very  serious. 

On  the  seacoast,  particularly  in  Southern  California,  heavy  salt 
fogs  cause  troubles  which,  as  far  as  the  speaker  knows,  have  not  yet 
been  overcome  successfully;  in  the  alkali  deserts,  the  so-called  salt 
storms  result  in  losses  by  very  remarkable  brush  discharges  and  leak- 
ages; and  where  lines  are  near  steam-railroad  right  of  way  the  oil  and 
water  from  the  exhaust  condense  on  the  insulators  and  then  collect 
coal  and  other  dust  until  the  creeping  surface  is  largely  covered,  caus- 
ing heavy  leakage,  and  burning  wooden  pins.  This  condition  promises 
to  be  a  very  serious  problem  in  steam-road  partial  electrification.  The 
deposits  from  salt  and  dust  storms  are  washed  off  by  the  rains,  but  the 
oily  coating  resulting  from  locomotive  exhausts  is  not  affected  by 
water. 

While  a  "campaign  of  education"  may  be  of  assistance,  the  small 

boy  with  his  sling-shot  and  the  man  with  the  gun  will  always  menace 

seriously  the  welfare  of  insulators  in  settled  sections.     Dark-colored 

glazes,  being  less  conspicuous,  are  being  used  in  many  cases.     With 

medium  voltages,  compact  insulators  of  the  Eedlands  or  Crown  type 

are  iised,  and  one  large  manufacturing  company  grooves  the  insulator 

top,  with  the  idea  that  the  marksman  will  knock  out  the  portion  inside 

the  groove  and  then  retire  satisfied,  leaving  enough  insulator  on  the 

pin  to  protect  the  line.     As  a  matter  of  fact,  against  a  bullet  of  any 

weight  there  is  little  choice  as  to  type.     The  speaker  tested  the  three 

38 
types  shown,  Fig.    1.    Plate    XXXV,    using    a   Winchester  r-rT,-caliber 

rifle,  reproducing  line  conditions  as  far  as  possible,  and  firing  one  shot 
at  each.  Pig.  2,  Plate  XXXV,  shows  the  result.  In  New  England 
or  the  East  generally,  a  gun  of  such  heavy  caliber  would  rarely,  if  ever, 
be  used. 

Serious  sleet  storms,  fortunately,  are  not  common,  and  are  rarely 
of  great  extent.  While  failure  may  be  due  to  the  dead  weight  of  the 
accumulation,  the  usual  cause  is  more  complex.  In  a  strong  wind,  the 
sleeted  wires,  because  of  the  greatly  increased  area  and  weight,  sway 


PLATE  XXXV. 

PAPERS,   AM.  SOC.   C.   E. 

MARCH,   1908. 

HARTE  ON 

OVERHEAD  CONSTRUCTION 

FOR    ELECTRIC    TRACTION. 


Fig.  1.— Insulators,  Before  the  Test. 


Fig.  3.— Insulators,  After  the  Test. 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  241 

until  the  vibrations  become  synchronous  with  the  natural  period  of  a  ^r.  Hartt 
pole  of  the  line.    If  this  pole  fails,  the  resulting  long  span  usually  has 
weight  enough  to  pull  down  other  poles  on  each  side. 

Aluminum  does  not  appear  to  hold  sleet  as  does  copper,  and  lines 
transmitting  much  power  are  usually  a  little  warmer  than  the  air,  so 
that  they  throw  oft"  the  coat  quickly,  therefore,  an  allowance  of  i  in. 
of  ice  around  the  wire  is  sufficient  to  provide  for  all  reasonable  con- 
tingencies. The  sleet  coatings,  however,  may  become  very  heavy.  Fig. 
1,  Plate  XXXVr,  shows  an  accretion  of  clear  ice  on  a  twig,  the  coat- 
ing having  a  diameter  of  practically  3  in. ;  and  C.  J.  H.  Woodbury, 
M.  Am.  Soc.  C.  E.,  recently  advised  the  speaker  of  specimens  he  had 
seen,  one  from  Wayland,  Mass.,  4J  in.  in  diameter,  and  one  from 
Nahant,  8  in.  in  diameter.  Fig.  2,  Plate  XXXVI,  shows  the  accumu- 
lation on  wires  caused  by  the  same  storm  which  developed  the  twig 
coating.     This  was  at  Winsted,  Conn.,  on  February  21st,  1898. 

Occasionally,  sleet  and  snow  storms  make  trouble  by  bridging  over 
the  creeping  surface  of  the  insulator,  thus  causing  leakage  to  the  cross- 
arm.  This,  however,  can  be  prevented  by  designing  the  pins  and  cross- 
arms  so  as  to  leave  no  large  catchment  area,  and  by  maintaining  a  con- 
siderable distance  from  the  insulator  top  to  the  arm. 

The  pressure  variations  in  wind  storms,  cited  by  Mr.  Coombs  as  a 
reason  for  using  a  low  value,  are  more  properly  reasons  for  introduc- 
ing a  sway  factor.  A  field  of  grain  or  long  grass,  or  ivy  on  a  house, 
observed  in  a  wind  storm,  shows  very  clearly  the  successive  pressure 
waves.  On  a  transmission  line,  when  such  impulses  are  synchronous 
with  the  natural  period  of  one  of  the  poles,  stresses  are  set  up  in  the 
latter  far  in  excess  of  those  due  to  the  direct  forces  themselves.  In 
wooden  pole  lines  there  are  stress  transfers,  due  to  the  elasticity  of 
the  poles  and  the  slipping  of  the  wires  at  the  insulators,  which  relieve 
these  unusual  conditions ;  in  tower  lines,  the  greater  rigidity  of  con- 
struction largely  prevents  such  relief,  and  the  action  miTst  be  con- 
sidered in  designing. 

Mr.  Coombs  gives  a  series  of  very  valuable  tables  of  wire  factors, 
but  there  is  a  matter  in  connection  with  solid  copper  that,  as  far  as 
the  speaker  is  aware,  has  been  given  practically  no  attention  by  any 
investigator.  The  standard  American  copper-wire  bar  weighs  approxi- 
mately 200  lb.  This  is  rolled  to  a  rod  of  diameter  depending  on  the 
gauge  of  the  wire  it  is  to  make.  For  0000  trolley,  this  rod  is  about 
560  mils  in  diameter,  and  is  less  than  300  ft.  long.  To  secure  the  long 
commercial  lengths  of  trolley,  the  rods  are  brazed  together,  the  scarf 
having  an  angle  of  about  20°,  the  brazing  being  done  with  a  mixture 
of  silver  and  tin  at  a  temperature  near  800°  fahr.  As  a  result,  the  rod 
is  annealed  at  the  scarf,  and  the  subsequent  drawings  to  a  diameter  of 
460  mils  do  not  harden  this  annealed  portion. 

From  tests  of  brazes,  it  appears  that  their  strength  is  only  eight- 


242 


DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 


Mr.  Harte.  tenths  of  that  of  stock  wire.  Incidentally,  it  should  be  noted  that  the 
strength  of  grooved  0000  wire  is  from  4  to  5%  lower  than  that  of 
round  wire,  due  to  the  fact  that  more  work  is  done  upon  the  latter. 
Grooved  wire  cannot  be  given  a  second  reduction  after  the  groove  has 
been  made. 


TABLE  11.— Tests  of  0000  Grooved  Trolley  Wire. 
Length  tested  in  each  case,  10  in. 
Tests  of  Brazed  Joints. 


to         M 

fl 

s 
p 

CO 
<D 

•s-g 

"S.S 

c3    . 
^  CI 

siS.S 
Cm 

W    5 

h^ 

m 

3  639 

0.475 

1.4 

5  852 

1.5 

3  630 

0.476 

1.5 

6  306 

6.7 

3  631 

0.448 

0.9 

4  429 

14.7 

3  632 

0.475 

1.4 

6  175 

6.5 

3  633 

0.452 

1.1 

i  515 

13.9 

3  634 

0.476 

1.3 

6  390 

3.8 

Average. 

0.467 

1.3 

5  611 

7.9 

Remarks. 


j  Apparently  good  braze.     Broke  in  joint,  with  par- 
1        tial  separation  of  braze. 

Apparently  erood  braze    Broke  outside  joint. 
\  Apparently  good  braze.    Metal  reduced  in  section 
-:         by  ttlirig.    Broke  in  joint,  with  partial  separa- 
(        tion  of  copper. 

Apparently  good  braze,  broke  outside  joint. 
^  Apparently  poor  braze.    Metal  reduced  in  section 
by  filing.      Broke  in  joint,  with  partial  separa- 
(        tion  of  copper. 

(  Apparently  good  braze.    Broke  in  joint.    Separa- 
tion very  slight.     Flaw  in  copper  at  point  of 
(        rupture. 


Tests  of  Wire  From  Sections  Between  Joints. 


3  628 

0.478 

7  087 

7.0 

3  635 

0.478 

7  085 

5.7 

3  636 

0.478 

7  166 

5.4 

3  637 

0.478 

7  077 

8.3 

3  638 

0.478 

.... 

7  076 

5.9 

Average. 

0.478 

7  098 

6.5 

Fracture  silky,  angular. 


The  failure  at  a  braze  does  not  occur  in  the  braze  itself,  but  in  the 
area  immediately  adjoining;  the  break  is  usually  parallel  to  the  scarf, 
but  there  is  invariably  a  skin  of  copper  on  the  braze. 

At  least  one  wire  manufacturer  uses  a  specially  heavy  wire  bar 
when  requested,  thus  having  fewer  brazes  per  mile,  but,  in  any  case, 
it  is  the  braze  which  determines  the  strength  of  the  line.  For  this 
reason,  as  well  as  because  of  its  flexibility  and  consequent  ease  of 
handling,  stranded  copper  is  much  better  than  solid  for  transmission- 


PLATE  XXXVI. 

PAPERS,   AM.  SOC.  C.   E. 

MARCH,   1908. 

HARTE  ON 

OVERHEAD  CONSTRUCTION 

FOR    ELECTRIC    TRACTION. 


Fig.  1.— Sleet  Accretion  on  Twig,  Winsted,  Conn.,  February  31st   1898. 


1  A/ 


,:?^i^^   ' 


Fig.  2.— Sleet  Accretion  on  Wires.  Winsted,  Conn.,  February  31st,  1898. 


Papers.]  DISCUSSION  ON  OVEEHEAD  ELECTRIC  TRACTION  243 

line  work,  the  strand  brazes  being  distributed  along  the  made-up  cable.  Mr.  Harte. 
It  should  be  added  that  the  problem  of  the  braze  is  now  occupying  the 
attention  of  a  number  of  the  large  wire  manufacturers,  and  it  is  hoped 
that  decided  improvements  will  follow. 

Mr.  Coombs  refers  to  a  grooved  trolley  wire  having  an  area  of 
0.155  sq.  in.  and  an  ultimate  strength  of  8  800  lb.  The  speaker  would 
like  to  have  further  details  of  this  wire.  Commercial,  American,  0000, 
grooved  wire,  having  a  cross-section  of  0.167  sq.  in.,  this  being  a  little 
more  than  7%  greater  than  the  wire  referred  to,  in  a  series  of  tests, 
failed  to  reach  an  ultimate  strength  of  8  000  lb.,  the  break  usually 
occurring  at  about  7  800  lb. 

Mr.  Coombs'  type  of  anchor  is  good  for  comparatively  light 
stresses,  but  for  heavy  spans  it  is  desirable  to  arrange  the  insulators  in 
pairs,  or,  if  in  tandem,  to  the  catenary  of  the  span,  to  secure  uniform 
stress  distribution. 

Where  the  sag  must  not  fall  below  fixed  limits,  provision  must  be 
made  for  adjustments  of  considerable  extent.  In  Mr.  Coombs'  design, 
any  considerable  take-up  on  the  turn-buckles  would  result  in  slack  on 
the  saddle,  not  readily  cared  for.  This  may  be  avoided  by  using  a 
double  saddle,  both  parts  being  movable,  the  slack  looping  between. 
The  Connecticut  River  crossing  of  the  Springfield-Suffield  line  has  a 
crossing  span  attached  to  a  movable  cross-head  controlled  by  a  long 
screw  with  an  adjusting  nut.  Fig.  9.  The  main  line  taps  into  the 
crossing  span  at  the  cross-head,  and  has  a  "pigtail"  to  care  for  the 
variation  in  length. 

Under  specifications,  to  bar  thin  wiped  galvanizing,  it  is  desirable 
tc  require  the  galvanized  metal  to  stand  four  immersions,  of  60  sec. 
each,  in  a  saturated  solution  of  copper  sulphate,  at  70°  fahr.  After 
each  immersion  the  test  piece  should  be  dipped  into  clean  water  and 
then  wiped  dry;  no  metallic  copper  should  appear  after  the  fourth 
immersion. 

The  speaker  wishes  Mr.  Coombs  had  treated  the  subject  of  protec- 
tion of  line  crossings  at  greater  length.  Apparently,  the  crossing  de- 
scribed is  protected,  over  and  above  the  general  line,  only  by  taking 
additional  precautions  to  relieve  the  crossing  towers  from  stresses  due 
to  adjoining  spans,  and  in  certain  details  of  anchorage  of  line,  provi- 
sion being  made  for  the  installation  of  a  cradle  at  a  later  date  if 
desired. 

As  far  as  line  strength  is  concerned^  a  crossing  differs  from  a  normal 
span  only  in  possible  greater  length,  or  in  restrictions  as  to  height  of 
wire;  and  a  failure  at  this  point,  in  its  effect  on  the  service,  does  not 
differ  from  a  failure  elsewhere;  but,  in  the  possibilities  of  damage  to 
train,  to  passengers  or  others  on  platforms  or  highways,  or  to  other 
lines,  the  crossing  becomes  one  of  the  most  critical  line  points,  and 
the  method  of  safeguarding  it  is  of  the  utmost  importance. 


244  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Harte.        Protection  may  be  effected  by : 

1. — Mechanically  preventing  a  broken  wire  from  getting  into 
the  danger  section; 

2. — Strengthening  the  upper  wires  so  that  failure  is  practically 
impossible; 

3.- — Grounded  arms  to  cut  ofi  a  broken  wire  at  the  pole  top  be- 
fore it  can  reach  the  line  below. 

Of  the  first  class  are  the  various  cradles,  all  open  to  the  criticism 
that  they  offer  large  areas  for  sleet  lodgment,  and  most  of  them  that 
they  are  very  expensive. 

A  very  simple  type  consists  of  telegraph  wires  strung  between  mul- 
tiple pin  arms.  The  system  is  grounded,  but  the  small  section  of  the 
wires  is  usually  a  guaranty  that  they  would  be  burned  through  by  the 
arc  in  case  of  a  fall  of  the  power  line.    Fig.  1,  Plate  XXXVII. 

A  modification  consists  of  three  longitudinal  wires  with  cross-bars 
of  hard  wood;  other  variations  include  sheets  of  wire  netting,  net- 
works of  wire  strand  more  or  less  substantially  fastened  together,  up 
to  the  very  impressive  cradles  of  heavy  strand  with  cross-bars  of  flat 
iron.  As  clearly  appears  in  the  case  shown.  Fig.  1,  Plate  XXXVIIl, 
wood  bar  cradles  are  apt  to  lose  members  from  breakage  or  otherwise, 
while  the  heavier  wire  cradles  often  sag,  becoming  a  positive  menace- 
In  Fig.  2,  Plate  XXXVII,  is  shown  a  wire-strand  cradle  which  has 
sagged  to  an  extent  requiring  the  power  company  to  protect  its  lines 
by  the  support  wires  strung  on  the  top  arm  of  the  transmission  line. 

While  the  more  substantial  types,  if  large  enough  to  keep  a  fallen 
wire  from  blowing  out  again,  are  no  doubt  efficient  along  certain  lines, 
their  great  cost,  and  the  excessive  stress  imposed  by  them  upon  their 
supports,  even  without  the  great  loads  of  sleet  they  are  sure  to  catch, 
make  them  very  undesirable. 

The  ideal  protection  is  the  so-called  short-span  method.  Here  the 
crossing  span  and  the  two  spans  adjoining  are  arranged  so  that  the 
distance  apart  of  the  poles  is  less  than  the  distance  from  the  cross-arm 
of  the  upper  line  to  the  lower  line;  it  is  thus  impossible  for  the  two 
lines  to  touch,  under  any  circumstances,  while  the  adjoining  short  spans 
prevent  a  broken  wire  from  swinging  into  the  crossing-span  section. 
Unfortunately,  crossings  usually  occur  in  highways  where  limitations 
as  to  pole  locations  prevent  the  use  of  this  method. 

The  method  most  generally  applicable,  and,  in  the  speaker's  judg- 
ment, the  best,  is  that  of  reinforcing  the  line  by  a  set  of  messenger 
cables.  This  plan  has  the  great  advantage  that  the  line  is  locally 
doubled  in  strength,  with  but  little  increase  in  weight  or  in  exposure 
area,  and  the  cost  is  nominal. 

In  a  recent  and  satisfactory  design,  messengers  of  No.  2  stranded 
copper  are  used,   and   to  them   the   line  wire   is  tied   every  4  ft.     A 


PLATE  XXXVII. 

PAPERS,   AM.  SOC.  C.  E. 

MARCH,  1908. 

HARTE  ON 

OVERHEAD  CONSTRUCTION 

FOR    ELECTRIC    TRACTION. 


Fig.  ].— Wire  Gridiron  Under  33  000- Volt  Transmission  Line. 


Fig.  2.— Wire  Cradle  Over  11  oOO-Volt  Transmission  Line 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  245 


Mr.  Harte. 


346  DISCUSSION"  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Harte.  grounded  angle-iron  frame  just  beneath  the  line  provides  an  automatic 
cut-o£F  in  case  the  line  breaks  in  the  span  adjoining  the  protected  span. 
Fig.  10,  and  Fig.  2,  Plate  XXXVIII. 

A  protection  device  should  meet  the  following  requirements: 

1. — Complete  protection  of  the  lower  line^  including  protection 
from  a  failure  in  an  adjoining  span  with  wires  whipping 
into  the  protected  section; 

2. — A  minimum  of  areas  exposed  to  sleet  or  wind; 

3. — Little  increase  over  weight  of  normal  line; 

4. — Simplicity  of  design; 

5. — Construction   familiar   to   linemen; 

6. — Few  special  parts; 

7. — Low  cost  of  installation  and  maintenance. 

The  foregoing  types  of  protection  assume  that  the  power  line  is 
above,  and  the  additional  safety  of  such  arrangement  will  usually 
justify  a  considerable  expenditure  to  secure  it.  In  some  cases,  how- 
ever, it  is  practically  impossible  to  go  above  with  the  transmission  line. 
In  such  cases  the  method  of  the  American  Telephone  and  Telegraph 
Company,  of  practically  enclosing  the  upper  line  in  a  sheath  of  |-in. 
strand  network  is  the  best,  provided  the  design  is  such  that  it  will 
prevent  dangerous  sagging  of  the  cradle. 

In  any  cradle  design,  the  tendency  of  a  broken  wire  to  curl  and 
therefore  jump  out  of  the  cradle  should  be  recognized ;  in  Germany  and 
Switzerland  it  is  customary  to  compel  transmission  companies  to  make 
all  lines  crossing  railroads  pass  through  a  regular  tunnel  of  ironwork. 

Where  the  line  is  on  very  narrow  right  of  way,  and  where  it  carries 
trolley  brackets,  straight  poles  are  essential,  and  it  is  often  desirable 
to  use  selected  stock  in  important  highways,  but  in  the  majority  of 
cases  considerable  crook  can  be  allowed.  Certainly,  where  chestnut 
is  to  be  used,  Mr.  Coombs'  requirement  of  only  1  in.  of  crook  in  10  ft. 
of  length  is  unnecessarily  rigid. 

Tbe  Western  Lumberman's  and  the  Idaho  Cedarmen's  Associations 
have  defined  commercially  straight  cedar  poles  as  having  a  crook  in 
one  direction  only,  and  a  sweep  not  to  exceed  1  in.  in  6  ft.  For  chest- 
nut, the  American  Telephone  and  Telegraph  Company  allows  practi- 
cally 1  in.  sweep  in  24  ft.  of  length  for  poles  up  to  40  ft.  total  length, 
and  of  1  in,  sweep  in  3  ft.  for  poles  more  than  40  ft.  in  total  length, 
the  measurements  to  be  made  between  the  top  and  a  point  6  ft.  from  the 
butt. 

Chestnut  from  seed  often  grows  very  straight;  stump-grown  stock, 
which  to-day  forms  a  large  proportion  of  the  supply,  almost  in- 
variably shows  a  sharp  crook  near  the  butt,  due  to  the  growth  of  the 
shoots,  first  out  to  clear  each  other  and  then  straight  upward.  If 
this  crook  is  large,  it  increases  the  cost  of  pole  setting,  but  a  diver- 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION 


24'^ 


Mr.  Harte. 


248  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Harte.  gence  from  the  general  axis  of  the  pole  of  not  more  than  12  in.  in  the 
lower  6  ft.  can  be  cared  for  without  additional  work. 

In  the  speaker's  judgment,  the  severity  of  the  specification  does  not 
increase  the  line  strength,  and,  as  it  materially  increases  the  cost,  it 
would  seem  that  it  might  better  be  changed,  to  meet  current  practice 
and  market  limitations,  to  the  following: 

Cedar  poles  shall  have  but  one  crook,  this  in  one  way  only,  the 
sweep  not  to  exceed  1  in.  in  6  ft. 

Chestnut  poles  shall  have  but  one  crook,  this  in  one  way  only. 
The  sweep  shall  not  exceed  the  following  limits  between  butt  and  top: 

Pole  length,   in  feet:  30—35—40—45—50—55—60—65—70. 
Sweep,    in    inches:         9— 10— 11— 11— 11— 12— 13— 14— 15. 

As  far  as  the  speaker  is  aware,  there  has  as  yet.  been  no  wreck  of 
any  magnitude  on  any  of  the  electrified  steam  lines,  and,  until  such  a 
try-out,  certain  questions  of  design  must  remain  imanswered. 

It  is  not  at  all  unlikely,  however,  that,  at  least  for  lines  on  which 
freight  trains,  with  their  capacity  for  trouble,  are  handled;  the  ulti- 
mate development  in  steam  railroad  electrification  will  be  in  the  di- 
rection of  independent  overhead  lines  for  each  track  or  group  of  tracks. 

With  one  live  trolley,  emergency  movements  can  be  made  on  ad- 
joining tracks;  with  all  overhead  wires  down,  as  may  well  be  feared 
in  a  wreck  under  bridge  construction,  not  only  is  the  electrical  equip- 
ment helpless,  but  a  large  additional  burden  of  clearing  away  the  ma- 
terial devolves  on  the  wrecker. 

Whatever  the  design,  the  speaker  feels  that  the  unit  stresses  allowed 
by  Mr.  Coombs  are  too  high.  With  long  and  frequent  trains,  and  par- 
ticularly with  high  voltages,  the  failure  of  any  part  of  the  overhead 
system  offers  too  great  an  opportunity  for  serious  results  to  justify 
any  close  paring  in  the  design. 

Steam  railroad  electrification  for  some  years  to  come  will  be  un- 
dertaken only  where  there  is  in  sight  a  very  marked  gain  by  the 
change,  or  where  legislation  compels  it.  The  complications  with  which 
the  simplest  distribution  system  involves  maintenance  operations,  and 
the  awkward  fact  that  for  wrecking,  and  in  ''dead"  sections,  some  self- 
contained  motor  must  be  used,  weigh  heavily  with  men  familiar  with 
steam-road  operation,  and  offset  many  of  the  obvious  advantages  of 
electrification.  It  will  rarely  happen  that  a  cost  variation  several  times 
in  excess  of  the  difference  between  thoroughly  dependable  construction 
and  "probably  safe"  construction  will  be  of  weight  in  influencing  the 
decision,  and  in  the  few  cases  where  it  is  a  factor  it  is  far  better  for 
the  art  that  the  work  be  deferred  rather  than  incur  an  unjust  discredit 
because  of  failure,  either  physical  or  in  performance,  as  to  expected 
maintenance  and  operation  costs. 

Whatever  the  unit  stresses,  the  bridges,  brackets,  or  poles  shouM 


PLATE  XXXVIII. 

PAPERS,   AM.  SOC.   C.   E. 

MARCH,   1908. 

HARTE    ON 

OVERHEAD  CONSTRUCTION 

FOR    ELECTRIC    TRACTION 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  249 

have  a  safety  factor  considerably  in  excess  of  that  of  the  overhead  Mr.  Harte. 
system  proper,  and  messengers,  hangers,  and  trolley  should  mark  a 
regiilarly  descending  scale,  in  order  that  any  failure  may  be  of  the 
least  extent  possible. 

While  it  is  trvie  that  the  ordinary  trolley  suspension  gives  a  catenary 
curve,  the  general  practice  to-day  is  to  apply  the  term  "'catenary  con- 
struction" to  systems  supporting  the  trolley  from  one  or  more  messen- 
gers.   Following  this  practice,  overhead  systems  wouW  be  classified  as: 

1. — Simple  Suspension. — Trolley  carried  by  hangers  directly 
connected  to  span  wire,  bracket,  or  bridge; 

2.— Catenary  Suspension. — Trolley  hung  from  one  or  more 
messenger  cables  in  turn  carried  by  the  span  wires, 
brackets,  or  bridges; 

3.- — Single  Catenary. — Having  but  one  main  messenger  (as  in 
the  Erie  Railroad  electriiication)  ; 

4.^Multiple  Catenary. — Having  more  than  one  main  messen- 
ger (as  in  the  New  Haven  electrification)  ; 

5. — Simple  Catenary.- — Trolley  carried  directly  by  main  messen- 
ger; may  be  simple  or  multiple  (as  in  the  Erie  or  New 
Haven  electrification)  ; 

6.— Compound  Catenary.— Trolley  carried  by  a  secondary 
messenger  system,  in  turn  carried  by  the  main  messenger 
(as  in  the  Blankenese-Ohlsdorf  Railway). 

The  kind  of  conductors  best  adapted  to  the  collection  of  the  power 
is  an  open  question.  The  two  chief  difficulties,  with  high  speeds,  are 
the  chattering  of  the  shoe,  due  to  alternate  hard  and  soft  spots  in  the 
line,  and  the  pressure  variations,  due  to  the  great  vertical  range  re- 
quired of  the  pantograph. 

The  first  problem  may  be  solved  either  by  floating  or  by  fixing  the 
trolley  wire;  for  undoubtedly  a  perfectly  flexible  line  or  one  perfectly 
rigid  would  give  excellent  results  as  far  as  it  alone  was  concerned. 
Whether  the  shoe  will  not  chatter  on  the  rigid  line,  as  a  result  of  the 
irregular  movements  of  the  car,  remains  to  be  seen.  It  is  interesting 
to  note  that  Mr.  W.  S.  Murray,  Electrical  Engineer  of  the  New  York, 
New  Haven,  and  Hartford  Railroad,  in  discussing  his  recent  paper 
before  the  American  Institute  of  Electrical  Engineers,  is  quoted*  as 
saying  that,  in  his  judgment,  either  the  shoe  or  the  line  must  be 
flexible. 

The  second  problem  is  largely  a  function  of  overhead  crossing 
limitations,  and,  to  a  large  degree,  is  independent  of  the  overhead 
construction;  therefore  it  must  be  cared  for  in  the  design  of  the  col- 
lector itself. 

Both  problems  are  of  the  field  rather  than  of  the  office.    Mr.  Mayerf 

*  street  Ruilicay  Journal.  January  18th,  1908,  page  8i. 
t  Proceedings,  Am.  Soc.  C.  E.,  for  December.  1907. 


250  DISCUSSION"  OK  OVERHEAD  ELECTRIC  TRACTION"  [Papers. 

Mr.  Harte.  has  given  a  very  elegant  mathematical  analysis  of  shoe  pressure  under 
certain  conditions,  but  the  discussion  is  based  on  the  supposition  that 
the  car  end  of  the  collector  traverses  a  path  bearing  a  definite  and 
regular  relation  to  the  conductor.  As  a  matter  of  fact,  however,  this 
path  is  most  irregular.  Unevenness  of  track,  as  to  grade  and  line, 
gauge  variations  of  rail  and  wheel,  side  play  in  axle  boxes,  spring 
action,  and  movements  in  the  car  framing  itself,  all  affect  the  shoe 
pressure  entirely  independently  of  the  variations  due  to  the  collector 
mechanism  and  the  character  of  the  overhead  system. 

Mr.  Coombs  recites  five  objections  to  the  double,  as  compared  with 
the  single,  catenary.  That  the  double  catenary  has  greater  first  cost 
and  greater  mass  overhead  is  true,  although,  by  the  time  the  single 
form  has  been  properly  secured  by  pull-offs,  guys,  and  steady  braces, 
there  is  a  surprising  amount  of  material  in  the  air. 

As  to  maintenance,  however,  the  speaker  doubts  whether  a  single 
catenary  is  not  at  least  as  troublesome.  A  double  catenary  can  stand 
severe  punishment  and  still  permit  the  movement  of  trains.  On  the 
other  hand,  the  hangers  of  the  single  catenary  are  more  out  of  the  way, 
and  therefore  less  likely  to  be  injured. 

Either  type  requires  the  tower  car  for  repairs,  but  the  double 
catenary  has  twice  as  many  connections  to  make;  on  the  other  hand, 
its  greater  strength  and  rigidity  undoubtedly  reduce  the  troubles  above 
the  trolley. 

That  the  double  catenary  offers  greater  obstruction  to  the  view  of 
the  signals,  the  speaker  cannot  admit.  If  the  signals  are  on  bridges, 
they  will  be  between  the  tracks,  and  a  curve  that  would  bring  the 
overhead  structure  across  the  line  of  sight  would  also  bring  the  pole, 
towers,  or  truss  posts  also  into  line.  The  difficulty  relates  to  the  sec- 
ondary supports  rather  than  to  the  type  of  suspension. 

Mr.  Coombs  sums  up  the  situation  admirably.  If  anything  remains 
to  be  said,  it  is  this :  In  the  present  state  of  the  art  there  is  a  great 
lack  of,  and  need  for,  data  resulting  from  practical  tests  of  the  various 
theories. 

In  closing,  the  speaker  wishes  to  express  his  obligations  to  the  many 
friends  who  have  kindly  assisted  in  the  experiments,  and  have  loaned 
illustrations  for  use  in  this  discussion. 
Mr.  Osgood.  Farley  Osgood,  Esq.  (by  letter). — If  the  high-tension  wires  are  of 
sufficient  mechanical  streng-th  to  have  a  factor  of  safety  of  3,  under 
correctly-assumed  general  conditions,  it  is  very  doubtful  if  a  con- 
ductor will  part  in  the  span. 

Up  to  crossings  of  600  ft.,  it  is  not  considered  that  the  wires  are 
likely  to  cross  in  high  winds,  even  though  spreaders  are  not  used,  as 
experience  seems  to  indicate  that  the  wires  will  swing  from  their  normal 
positions  about  equally. 

Protection,  in  the  form  of  lightning  rods,  seems  desirable  at  cross- 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  251 

ings  where  very  high  wooden  towers  are  used,  or  on  lower  wooden  cross-  Mr.  Osgoori. 
ings  at  points  of  high  altitude. 

If  steel  towers  are  used  at  railroad  crossings,  the  use  of  lightning 
rods  is  desirable,  if  the  crossings  are  at  such  points  in  the  line  as  are 
known  to  he  affected  by  lightning  disturbance. 

The  use  of  cradles,  suspended  from  high-tension  poles,  under  the 
high-tension  wires,  is  not  advocated  by  the  writer,  for  any  ordinary 
circumstances. 

An  ideal  high-tension  crossing  would  have  the  supporting  towers  of 
sufficient  height  to  make  it  impossible  for  one  of  the  transmission  wires 
to  touch  the  ground  in  case  it  should  break  in  the  crossing  span,  but 
this  condition  is  usually  impossible,  from  a  rational  standpoint,  owing 
to  the  length  of  the  section. 

A  second  choice  for  crossings  seems  to  favor  a  supporting  tower  at 
the  edge  of  the  railroad  company's  right  of  way,  and  another  between 
the  telegraph  or  signal  wires  and  the  outside  rail,  thus  using  four  poles 
or  towers  to  a  crossing,  making  a  short  span  on  each  side  of  the  tracks 
over  the  telegraph  or  signal  wires  and  the  longer  span  over  the  tracks, 
it  being  assumed  that  the  railroad  company  has  wires  to  be  protected 
on  each  side  of  its  tracks. 

These  poles  should  be  of  sufficient  height  to  prevent  one  of  the 
high-tension  conductors  from  touching  the  telegraph  or  signal  wires, 
if  it  should  break,  so  that,  if  a  high-tension  conductor  should  give  way 
over  the  tracks,  the  signaling  system  would  not  be  affected,  although 
this  might  not  always  be  true  if  any  of  the  rails  were  used  as  part  of 
the  signaling  circuit. 

If  railroad  companies  feel  that  screens  or  cradles  should  be  placed 
under  high-tension  wires,  let  such  devices  be  placed  in  the  top  arm 
positions  on  the  poles  carrying  the  telegraph  or  signal  wires,  as  their 
purpose  can  then  be  accomplished  and  no  unnecessary  burden  be  placed 
on  the  more  important  high-tension  towers. 

A  simple  and  inexpensive  type  of  screen,  if  used  as  suggested,  can 
be  made  up  as  follows:  In  the  top  e?in  cf  the  pole  on  each  side  of  the 
high-tension  crossing,  place  a  ten-pin  cross-arm,  somewhat  longer  than 
the  standard  cross-arm  used  on  the  line  to  be  protected,  and  run  in 
between  these  arms,  ten  No..  6  or  No.  8  galvanized  steel  wires,  strapped 
together,  and  grounded  at  each  pole. 

W.  S.  Murray,  Esq.^  (by  letter). — Mr.  Coombs  has  handled  this  very  Mr.  Murray, 
interesting  subject  in  an  analytical  and  conservative  manner.  With 
reference  to  that  part  of  his  paper  concerning  the  strength  of  the  ma- 
terials to  be  used  for  wires  and  supporting  structures,  the  basis  of  his 
assumptions  could  not  be  better  founded  than  on  the  records  given  in 
liis  several  tables.  With  reference  to  the  equations  relating  to  the  unit 
pressures  per  square  foot  of  projected  area,  the  writer  is  pleased  to  be 

*  Electrical  Engineer,  New  York,  New  Haven  and  Hartford  Railroad. 


252  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  [Papers. 

Mr.  Murray,  able  to  confirm  these  figures  in  actual  practice,  as  the  catenary  wires 
and  supporting  structures  in  the  New  Haven  electrification  were  worked 
out  by  an  equation  practically  identical  with  the  one  suggested  by 
Mr.  Coombs;  and  it  is  of  interest  to  note  here  that  these  wires  and 
structures  have  passed  through  storms  approximating  quite  closely 
those  stated  as  maximum  conditions  upon  which  the  equations  are 
based. 

Mr.  Coombs'  specifications  of  general  requirements  are  to  the  point, 
and,  in  addition  to  those  which  are  generally  recognized  as  standard, 
he  has  made  many  original  and  valuable  suggestions. 

In  connection  with  the  general  subdivision  of  superstructures,  al- 
though the  writer  is  not  quite  able  to  agree  with  Mr.  Coombs  that  the 
upright  signals  when  supported  from  four-track  trusses  are  obscured 
from  the  engineer's  view  at  a  distance  of  1  200  ft.,  it  is  unquestionably 
true  that  the  general  envelope  produces  a  difiicult  foreground  for  the 
engineer,  and  naturally  the  cross-span  or  cantilever-bracket  construc- 
tion clears  up  this  disadvantage  to  a  considerable  degree. 

As  recently  stated  in  a  paper  before  the  American  Institute  of 
Electrical  Engineers,  the  writer  is  not  loathe  to  believe  that  even  four- 
track  main-line  electrification  will  be  effected  by  the  use  of  cross- 
catenary  spans  interspersed  at  proper  intervals  with  fabricated  steel 
truss  anchor  bridges;  but  believes  that  the  form  of  this  construction 
will  be  guyed  steel  uprights  siipporting  the  cross-catenary  span,  with 
distances  between  bents  of,  say,  not  greater  than  300  ft.;  and,  further, 
he  believes  that,  in  the  future,  the  single  catenary  will  receive  more 
favorable  consideration  than  the  double  catenary  construction.  It  can 
be  readily  seen  that  the  first  cost  of  the  former  will  be  much  less,  and 
the  flexible  contact  offered  by  the  single  catenary  construction,  due  to 
the  fact  that  the  trolley  is  supported  from  a  single  messenger,  with  the 
messenger  in  turn  supported  from  a  flexible  cross-catenary,  gives  it  a 
great  advantage. 

Practice  seems  to  demonstrate  the  fact  that  either  the  shoe  or  the 
trolley  must  be  flexible.  As  a  matter  of  fact,  flexibility  in  both  would 
be  of  great  advantage,  and  it  cannot  be  questioned  that  the  cross- 
catenary  span  will  offer  more  flexibility  than  either  the  cantilever  or 
bridge-truss  type  of  construction.  At  this  point,  particular  attention 
is  called  to  the  fact  that  experimentation  with  the  deflection  of  trolley 
wire  supported  from  a  messenger,  which  is  in  turn  supported  at  rigid 
points,  shows  that  in  the  middle  of  the  span  the  deflection  is  as  much 
as  400%  greater  than  that  in  the  immediate  vicinity  of  the  bridge  or 
cantilever  supporting  the  messenger  wire,  it  being  understood,  of 
course,  that  equal  upward  pressures  are  applied  in  each  instance.  This 
illustrates  the  value  of  the  flexible  feature  in  the  cross-catenary  support. 
A  point  of  much  value  in  the  cross-catenary  construction  should  be 
emphasized,  namely,  that  the  cross-spans  may  be  supported  on  strain 


Papers.]  DISCUSSION  ON  OVERHEAD  ELECTRIC  TRACTION  253 

insulators,  thereby  not  only  doubling  the  actual  insulating  value  of  the  Mr.  Murray. 
line,  as  measured  under  normal  atmospheric  conditions,  but,  in  point 
of  fact,  many  times  increasing  the  insulating  value  due  to  the  insula- 
tion being  placed  at  the  side,  and  thus  out  of  the  direct  line  of  steam 
locomotive  blasts,  which  have  such  a  deleterious  effect  on  insulation. 

An  argument  that  will  be  advanced  against  the  use  of  the  cross- 
catenary  construction  is  that  it  is  not  as  reliable  as  the  cantilever  or 
truss  construction.  The  answer  to  this  is  that,  in  this  form  of  con- 
struction, any  factor  of  safety  that  may  be  used  in  other  types  can  be 
selected;  in  fact,  larger  factors  of  safety  can  be  chosen  with  less  pro- 
portionate expense. 

In  conclusion,  the  writer  agrees  with  Mr.  Coombs  in  his  summa- 
tion, under  five  counts,  concerning  the  undesirability  of  double  cate- 
naries. The  root  of  all  trouble  with  the  alignment  of  catenary  con- 
struction is  the  change  of  temperature.  The  fact  that  a  low  tempera - 
tiire  means  a  tight  wire  and  vice  versa  for  a  high  temperature  must 
be  considered.  The  ideal  condition  of  suspension  would  be  a  free- 
running  suspended  wire,  tension  being  supplied  at  one  or  both  ends 
to  counteract  the  variations  in.  its  length  due  to  temperature.  It  is 
very  seldom  that  ideal  conditions  can  be  secured  in  the  field,  however, 
and  the  results  are  generally  a  combination  of  compromises  and  ap- 
proximations. What  one  fails  to  accomplish  with  the  contact  wire 
may  be  accomplished  by  a  properly  devised  shoe,  of  strong  construc- 
tion, flexible  and  light,  the  last-named  element  eliminating  inertia, 
the  arch  enemy  to  the  hard  spots  in  the  line,  which,  as  Mr.  Coombs  has 
pointed  out,  are  at  the  "hanger  points."  To-day  is  not  the  time  for 
standardization,  but  for  observation.  The  experiences  and  mistakes  of 
to-day  will  be  invaluable  in  comparison  with  theories. 


Vol.  XXXIV.  MARCH,  1908.  No.  3. 


AMEEICAN  SOCIETY  OF  CIVIL  ENaiNEERS. 

INSTITUTED    1852. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


A  NEAV  SUSPENSION  FOR  THE  CONTACT  WIRES  OF 
ELECTRIC  RAILWAYS  USING  SLIDING  BOWS. 

Discussion.* 


By  Messrs.  R.  D.  Coombs  and  Charles  Eufus  Harte. 


Mr.  Coombs.  R.  D.  CooMBS,  M.  Am.  Soc.  C.  E. — The  speaker  is  of  the  opinion 
that  interruptions  in  service  caused  by  lateral  displacement  are  im- 
probable on  either  the  240-ft.  spans  with  a  sag  of  2^  ft.,  as  used  by 
Mr.  Mayer,  or  on  300-ft.  catenary  spans  having  a  sag  of  6  ft. 

Based  on  the  fact  that  the  hot  and  cold  sags  in  the  catenary,  and 
therefore  in  the  trolley  wires,  are  approximately  equal  for  a  total  varia- 
tion  in  temperature  of  140°,  the  tension  in  the  trolley  wire  is  given 
by  Mr.  Mayer  as  about  26  000  lb.  per  sq.  in.  Assuming  that  the  cate- 
naries are  erected  with  the  normal  tension  at  normal  temperature,  it 
would  seem  that  the  increased  tension  in  the  trolley  wire  should  be 
merely  that  due  to  a  rise  or  fall  in  temperature  of  half  the  total  varia- 
tion. 

The  speaker  is  not  familiar  with  the  details  of  the  automatic  ad- 
justment of  the  trolley  wire  used  in  the  Blankenese-Ohlsdorf  line,  or 
other  foreign  lines  equipped  with  the  secondary  catenary,  but  thinks 
it  should  not  be  necessary  to  run  the  comparatively  inflexible  trolley 
wire  over  the  adjusting  pulleys,  as  this  might  be  avoided  by  attaching 
a  flexible  wire  which  would  permit  the  use  of  pulleys  of  moderate  diam- 
eter. 

The  elastic  limit,  of  from  40  000  to  45  000  lb.,  assumed  for  trolley 
wire  having  an  ultimate  strength  of  from  60  000  to  60  000  lb.  per  sq. 
in.  seems  to  be  rather  high,  and  the  maximum  stress  of  about  26  000  lb. 

*  This  discussion  (of  the  paper  by  Joseph  Mayer.  M.  Am.  Soc.  C.  E.,  printed  in 
Proceedings  for  December,  1907),  i.s  printed  in  Proceedings  in  order  that  the  views 
expressed  may  be  brought  before  all  members  for  further  discussion. 


PLATE   XXXIX. 

PAPERS,  AM.  SOC.  C.   E. 

MARCH,   1908. 

HARTE  ON 

SUSPENSION   FOR  WIRES 

OF  ELECTRIC  RAILWAYS. 


Improvised  -  Atwood's  Machine  "  for  Testing  the  Stiffness  of  Trolley  Wire. 


Papers.]   DISCUSSION"  ON  SUSPENSION  FOR  ELECTRIC  RAILWAYS  255 

per  sq.  in.  in  tlie  trolley  Avire  under  dead  load,  plus  6.0  lb.  wind  pres-  Mr.  Coombs, 
sure,  plus  bending  of  the  wire,  does  not  give  the  contact  wire  in  the 
saddle  suspension  the  same  factor  of  safety  as  that  used  in  designing 
first-class  railroad  bridges. 

A  parallel  condition  to  that  of  many  railroad  bridges  would  be  dead 
load,  plus  ice  ^  in.  thick,  plus  8  lb.  per  sq.  ft.  for  wind  pressure,  plus 
lateral  bending;  the  total  stress  from  which  would  exceed  26  000  lb. 

Assuming  that  the  single  catenary  and  the  trolley  wire  supported 
by  it  can  be  designed  with  suitable  factors  of  safety,  and  constructed 
so  as  to  give  satisfactory  operation,  the  extra  expense  of  the  messenger 
wire  and  hangers  may  be  justified  as  a  safeguard  to  prevent  falling 
v/ires  and  the  troubles  incident  to  them. 

Charles  Eufus  Harte,  M.  Am.  Soc.  C.  E. — Mr.  Mayer  has  de-  Mr.  Harte. 
veloped  a  very  interesting  construction  which  it  is  to  be  hoped  may 
have  a  practical  trial  in  the  near  future.  At  the  same  time,  it  should 
be  noted  that  the  excessive  stresses  feared  by  Mr.  Mayer  do  not  always 
develop  in  the  older  forms  of  suspension.  Undoubtedly,  on  long  level 
tangents,  with  heavy  anchoring,  there  would  be  heavy  stresses  at  low 
temperatures,  if  the  trolley  had  been  well  pulled  up  in  warm  weather, 
but,  as  a  matter  of  fact,  grade  changes  and  curves  offer  relief,  and 
trolley  pulled  to  a  tension  of  2  200  lb.  in  summer  apparently  does  not 
materially  increase  this  stress  in  winter  imder  usual  conditions,  owing 
to  the  yielding  of  supports.  Where  trolley  is  hung  slack,  however,  the 
changes  of  length  are  chiefly  taken  up  in  the  sag,  and  here  an  adjuster 
may  be  desirable;  but,  certainly  in  New  England,  such  a  device  must 
be  automatic  or  else  receive  constant  attention  in  order  to  meet  the 
rapid  and  large  changes  of  temperature  of  that  climate. 

A  system  of  counterweights  offers  the  ideal  method  of  securing 
uniform  tension,  and  there  is  little  difficulty  in  arranging  bell-cranks 
or  of  splicing  into  the  trolley  a  section  of  steel  strand,  if  the  trolley 
itself  is  too  stiff  to  lead  direct  to  the  counterweights.  It  must  be  ap- 
parent, however,  to  anyone  familiar  with  trolley  wire,  that  Mr.  Mayer's 
figures  and  practical  conditions  do  not  agree. 

To  determine  roughly  the  flexibility  of  0000  B.  &  S.  gauge  hard- 
drawn,  grooved  copper  trolley  wire,  the  speaker  arranged  a  crude  form 
of  the  "Atwood's  machine,"  of  physics,  the  trolley  wire  forming  the  cord, 
and  the  head  sheaves  of  a  dumb-waiter  the  wheel.  (Plate  XXXIX.) 
Balanced  weights  were  hung  from  the  wire,  and  then  one  side  was 
loaded  until  motion  occurred.  No  correction  was  made  for  the  con- 
siderable friction  of  the  wheels  used. 

With  a  pidley  65  in.  in  diameter  at  the  root  of  the  groove :  127  lb. 
on  each  side  required  13  lb.  additional  on  one  side  to  move;  232  lb. 
on  each  side  required  18  lb.  additional  on  one  side  to  move;  and,  792 
lb.  on  each  side  required  64  lb.  additional  on  one  side  to  move. 


256  DISCUSSION  ON  SUSPENSION  FOR  ELECTRIC  RAILWAYS  [Papers. 

Mr.  Haite.        With  a  wheel  33  in.  in  diameter :  232  lb.  on  each  side  required  33 
lb.  additional  on  one  side  to  move. 

With  a  wheel  17  in.  in  diameter:  127  lb.  on  each  side  required  83 
lb.  additional  on  one  side  to  move;  and,  324  lb.  on  each  side  reqiiired 
155  lb.  additional  on  one  side  to  move. 

The  Blankenese-Ohlsdorf  trolley  is  described*  as  a  grooved  wire, 
having  a  gauge  practically  equivalent  to  0000;  Mr.  Mayer  gives  the 
area  as  100  sq.  mm.,  which  is  3|%  less  than  the  area  of  the  wire  used 
in  the  foregoing  rough  test. 

For  assistance  in  the  tests,  the  speaker  is  greatly  indebted  to  his 
assistant,  Mr.  John  F.  Trumbull,  and  to  Messrs.  0.  W.  Blakeslee  and 
Sons,  Contractors,  of  New  Haven. 

*  Street  Railway  Journal,  April  6,  1907. 


Vol.  XXXIV.  MARCH,  1908.  No.  3. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEEES. 

INSTITUTED     1852. 


PAPERS  AND   DISCUSSIONS. 

Tliis  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  jjublications. 


SAFE  STRESSES  IN  STEEL  COLUMNS. 

Discussion.* 


By   Messrs.    Henry   B.    Seaman,   Luzerne    S.    Cowles,    Charles   M. 

Emmons,    Henry    S.    Prichard,    Horace    E.    Horton,   F.    P. 

Shearwood,  L.  D.  Rights,  and  A.  W.  Carpenter. 


Henry  B.  Seaman,  M.  Am.  Soc.  C.  E.  (by  letter). — It  may  be  too  Mr.  Seaman, 
early  yet  for  a  Special  Committee  to  advise  as  to  the  proper  column' 
formula  to  be   used  in  structural  work,   but   Mr.   Worcester's   paper 
brings  us  one  step  nearer  its  appointment. 

To  the  writer's  mind,  there  never  has  been  sufficient  reason  for 
abandoning  the  Rankine  formula.  The  basis  of  its  formation  is  the 
provision  that  a  column  receives  both  direct  strain  and  bending  strain. 
The  direct  strain  is  readily  provided  for,  and  the  effect  of  bending  is 
found  by  experiment,  the  results  of  which  are  used  in  determining  the 
coefficient  of  r^.  It  would  seem  better  to  plot  the  results  of  these  tests 
upon  the  basis  of  ultimate  strength,  rather  than  working  strength,  as 
it  keeps  the  mind  more  directly  on  the  actual  data  observed.  The 
formula  can  then  be  modified  for  working  strength,  either  by  taking  a 
certain  proportion  of  the  numerator  as  a  factor,  or  by  other  modifica- 
tion, if  preferred. 

It  should  be  remembered  that  details  are  designed  upon  an  assumed 

value  of  ,  =  12,  where  I  equals  the  length  and  d  the  least  diameter  of 

a  solid  rectangular  column.  In  designing  columns,  therefore,  a  greater 
strain  should  not  be  permitted  than  that  for  which  the  details  are  de- 
signed, that  is,  for  a  less  value  of   ,    than   12.     This  serves,   as  Mr. 

•  This  discussion  (of  the  paper  by  J.  R.  Worcester,  M.  Am.  Soc.  C.  E.,  printed  in  Pro- 
ceecUnc/s  for  January,  1908),  is  printed  in  Proceedings  in  order  that  the  views  expressed 
may  be  brought  before  all  members  for  further  discussion. 


258  DISCUSSION  OK  SAFE   STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Seaman.  Worcester  has  expressed  it,  to  truncate  the  formula  for  very  short 
columns.  Since  the  Rankine  formula,  however,  provides  for  bending 
as  well  as  compression,  it  conforms  with  the  tests  on  long  columns  bet- 
ter than  does  the  straight-line  formula,  and,  for  that  reason,  to  the 
v.'riter's  mind,  is  the  most  valuable  formula  we  have.  There  has  never 
seemed  to  be  any  excuse  for  the  adoption  of  a  straight-line  formula, 
except  simplicity  in  plotting  and  ease  in  memorizing.  Confessedly,  it 
does  not  conform  to  tests  on  long  columns,  it  is  not  applicable  beyond 
certain  restricted  limits,  and  finally,  since  it  involves  r  instead  of  r'^j 
it  cannot  be  used  as  readily  without  the  assistance  of  tables;  yet  tables 
will  assist  equally  well  with  any  formula. 

A  recent  study  of  the  tests  mentioned  by  Mr.  Worcester  has  led 
the  writer  to  adopt  the  following  fonnulas : 

For  Steel  and  Wrought  Iron:  For  Cast  Iron: 


8  000  7-2  1  000  r^ 

Mr.  Worcester  very  properly  calls  attention  to  the  fact  that  the 
failure  of  a  column  occurs  when  it  begins  to  cripple,  while,  with  the 
tension  member,  if  allowed  time  to  rest,  the  material  becomes  even 
stronger  because  of  the  work  of  overstrain  which  it  has  received.  This 
would  enable  us  to  permit  a  higher  factor  of  safety  upon  tension  mem- 
bers than  upon  compression  members,  were  it  not  for  the  fact  that  a 
permanent  elongation  of  a  tension  member  would  deform  the  struc- 
ture to  such  an  extent  as  to  change  the  strains  for  which  it  was  de- 
signed, and  possibly  cause  failure  on  that  account.  It  must  also  be 
remembered  that  the  element  of  fatigiae — and  possibly  that  of  moment- 
ary impact — need  not  be  considered  in  the  bending  of  the  column, 
and  therefore  the  extra  material  used,  in  order  to  prevent  bending,  is 
an  additional  factor  of  safety,  which  the  tension  member  does  not 
possess. 

The  recent  tendency  in  structural  design  seems  to  be  to  increase 
the  live  loading  by  a  given  factor  in  order  to  derive  an  equivalent 
static  strain,  and  then  to  design  the  parts  for  these  static  strains, 
rather  than  the  old  method  of  using  a  factor  of  safety  to  cover  defects 
in  material,  increase,  and  extraordinary  effects  of  loading,  etc.  If 
the  live-load  strains  can  be  increased  so  as  to  cover  all  possible  con- 
tingencies, and  if  a  dead  load  can  be  assumed  which  will  not  be  ex- 
ceeded under  any  circumstances,  it  would  seem  safe  to  place  the 
allowable  strain  at  one-half  or  two-thirds  of  the  elastic  limit.  It  is 
on  this  basis  that  long-span  bridges  are  designed ;  and,  by  the  adaption 
of  a  formula  in  which  this  factor  would  vary  with  the  various  lengths 
of  span,  the  same  method  of  proportioning  could  be  adopted  for  shorter 
spans.  Future  specifications  will  probably  tend  in  tlu'  direelion  of  some 
such  method  of  design. 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS  259 

Luzerne  S.  Cowles,  Assoc.  M.  Am.  See.  C.  E.  (by  letter). — Mr.  Mr.  Cowies. 
Worcester  need  hardly  offer  an  apology  for  continuing  the  agitation  con- 
cerning compressive  stresses  to  be  allowed  in  designing  structural  steel- 
work. The  writer,  from  the  beginning  of  his  career,  and  in  fact  during  his 
college  course,  has  been  decidedly  baffled  by  the  numerous  formulas  for 
allowable  safe  compressive  stresses,  and  had  begun  to  believe  that  most 
so-called  "rational"  formulas  were  made  up  to  assist  the  designer  in 
making  a  comparatively  safe  gxiess.  The  question  arises,  however,  as 
to  whether  all  the  commonly  accepted  formulas  do  really  give  the 
margin  of  safety  that  is  desired  and  is  assumed  to  exist. 

C.  C.  Schneider,  Past-President,  Am.  Soc.  C.  E.,  has  frequently 
called  attention  to  the  fact  that  the  elastic  limit,  and  not  the  ultimate 
strength,  should  be  especially  considered  in  deciding  the  real  factor 
of  safety.  This  gives,  for  tension,  and  supposedly  for  compression, 
a  real  factor  of  approximately  2,  on  the  basis  of  16  000  lb.  per  sq.  in. 
for  static  loads.  The  writer  agrees  with  this,  particularly  where  com- 
pression is  involved. 

In  the  light  of  recently  published  data  of  experiments  on  full-sized 
compression  members,  it  would  seem  that  this  real  factor  of  safety  of 
2  had  even  been  seriously  encroached  upon,  leaving  far  too  lean  a 
margin  of  safety  for  structures  where  human  life  is  at  stake.  When 
one  considers  the  astounding  results  of  Mr.  Buchanan's  tests,*  where 
the  fiber  stress  at  crippling,  even  for  so-called  "short"  columns,  was 
below  the  accepted  elastic  limit,  it  seems  to  be  high  time  to  consider 
reducing  the  allowable  unit  stress  for  compression  below  that  for  ten- 
sion, even  though  the  modulus  of  elasticity  and  the  elastic  limit  appear 
in  the  laboratory,  and  no  doubt  are,  approximately  the  same  for  each. 

Most  railroad  bridge  specifications  insist  that  no  compression  mem- 
ber shall  have  a  length  exceeding  100  times  its  least  radius  of  gyration, 
except  for  bracing,  where  a  ratio  of  120  may  be  used.     In  other  words, 

a  main  compression  member  in  which  the    -  is  100,  will  carry  safely 

f)  000  lb.  per  sq.  in.,  whereas  the  use  of  a  main  member  in  which  the 

-  is  greater  than  100  is  disapproved.     This  is  according  to  a  standard 

straight-line  formula,  and  it  seems  that  the  use  of  very  "long"  columns 
is  not  discouraged  to  the  extent  that  it  should  be. 

Is  not  then  the  really  "rational"  formula  one  which  gives  com- 
paratively low  results  for  the  allowable  fiber  stress  for  the  longest 
columns  consistent  with  good  design,  and  errs  on  the  side  of  safety 
for  the  occasional  exceptionally  short  strut  ?  Mr.  Worcester's  proposed 
formula  seems  to  fill  these  conditions,  and  while  it  may  not  be  perfect 
in  its  present  .form,  it  is  surely  a  step  in  the  right  direction,  and  fur- 
nishes   a   basis    for    a   truly    sensible   formula.      With    his    customary 

*Engineering  News,  Vol.  LVIIT,  pp.  685-695. 


2 GO  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  cowies.  modesty,  the  author  of  the  proposed  formula  has  failed  to  point  out 
its  commendable  features.    The  writer  suggests  the  following: 

(a). — Eeducing  the  allowable  stress  for  "short"  columns  so  as  to 
give  a  reasonable  factor  of  safety; 

(b). — Discouraging  the  use  of  columns  in  which  the      is  greater 
than,  say,  90  to  100; 

(c). — Placing  the  allowable  stresses,  for  columns  in  which  the 

I 
ratios  of       lie  between  30  and  90,  at  figures  which  are 

slightly  below  the  average  results,  as  shown  by  numerous 
tests. 

Charles  M.  Emmons,  M.  Am.  Soc.  C.  E.  (by  letter). — The  writer 
is  much  interested  in  this  paper.  Mr.  Worcester's  plotting  of  the  re- 
sults of  actual  tests,  reduced  by  a  safe  working  factor,  and  also  his 
plotting  of  the  several  column  formulas,  to  the  same  scale,  reveals  very 
graphically  the  inconsistencies  and  the  wide  divergencies  of  these 
formulas. 

The  writer  is  not  as  fully  impressed  with  the  idea  of  a   formula 

being  self-limiting  at  the  highest  allowed  value  of     •     In  attempting 

to  do  that,  the  author's  curve  appears  to  be  as  inconsistent  with  the 
tests  as  would  be  a  straight-line  formula.     The  writer  realizes  that  a 

formula  should  be  of  such  form  that,  if  the  allowed  value  of       be  not 

r 

fixed  arbitrarily,  it  will  yet  be  in  no  wise  dangerous,  for  the  reason 
that  someone  with  more  "nerve"  than  judgment,  or  through  ignorance 

or  other  cause,  will  occasionally  use  such  a  formula  as  


^  3()000r- 


to  the  limit.  This  danger  was  just  lately  brought  to  the  writer's  atten- 
tion in  a  case  where,  for  compression  members,  more  than  10  ft.  long, 
having  a  stress  of  about  4  000  lb.,  a  prominent  engineer  used  a  single 

I 
angle  2^-  by  2  by  i  in.     The      is  more  than  300,  and  yet,  according  to 

the  formula,  it  should  carry  the  load  safely. 

The  use  of  any  formula  which  may  be  based  on  a  series  of  observa- 
tions, like  those  given  by  the  author,  should  not,  with  confidence,  be 
])ushed  very  far  beyond  the  limits  of  those  observations.  Such  a 
formiila,  however,  should  take  full  advantage  of  what  is  indicated  as 
safe  by  those  observations. 

Again,  the  formula  proposed   by  the   author,  where  one  would  be 

practically  limited  to  -  =  112,  would  be  prohibitive,  in  many  classes 
of  work,  especially  for  secondary  members. 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS  261 

In  view  of  these  considerations,  the  writer  would  prefer  the  use  of  Mr.  Emmons. 
a  parabola  with  values,  say,  12  000  (1  —    "f^n  7x2  )  '  ^^^^^  ^^^'  limiting 

I 
allowed  value  of       prescribed.     The  formula  is  practically  self -limit- 
ing  at       =  150.     It  will  be  observed  that  this  curve  gives  practically 

the  same  values  as  the  circle  up  to       =  80,  and  from  that  point  it 

follows  the  tests  far  better,  taking  advantage  of  what  the  tests  indicate 
as  safe,  and  yet  in  no  case  becoming  dangerous. 

Henry  S.  Prichard,  M.  Am.  Soc.  C.  E.  (by  letter). — The  intro- Mr.  Pricimrd. 
ductory  paragraphs  of  this  paper  give  the  impression  that  the  author 
attaches  slight  importance  to  theory  in  regard  to  columns.  Would  it 
not  be  well  to  discriminate  somewhat  in  this  regard  ?  It  is  unfortunate 
that  a  single  word,  "Theory,"  is  popularly  used  (with  the  sanction  of 
the  dictionaries)  to  designate  "a  body  of  the  fundamental  principles 
underlying  any  science  or  application  of  a  science,"  and  the  radically 
different  conception  "a  proposed  explanation  designated  to  account  for 
any  phenomena,"  no  matter  how  visionary  the  assumptions,  fallacious 
the  argument,  or  foolish  the  conclusion.  It  is  natural  and  proper  that 
many  of  the  proposed  explanations  of  the  behavior  of  columns  should 
be  held  in  light  esteem,  but  it  is  highly  desirable  that  engineers  should 
understand  and  apply  the  principles  of  mechanics  to  the  design  of 
columns.  Without  such  an  understanding,  the  phenomena  observed 
in  practice  and  in  the  numerous  compression  tests  are  to  a  considerable 
extent  a  set  of  seemingly  discordant  facts. 

Referring  to  the  fact  that  the  practice  of  steel  designers  with  re- 
gard to  columns  may  well  bear  further  consideration,  the  author  states: 

"The  reasor^  for  this  is  that  all  'rational'  column  formulas,  based 
on  the  elastic  properties  of  steel,  are  founded  on  considerations  which 
are  applicable  only  to  ratios  of  length  to  radius  of  gyration  far  beyond 
those  allowed  in  actual  construction." 

It  is  difficult  to  reconcile  this  statement  with  the  analyses  and 
equations  developed  by  Euler,  Cain,  Fidler,  Marston,  J,  B.  Johnson, 
Moncrieff,  and  others  who  have  determined  important  facts  regarding 
short  as  well  as  long  columns  by  reasoning  based  on  the  elastic  prop- 
erties of  steel  and  iron.  The  names  of  Tredgold,  Gordon,  and  Rankine 
have  purposely  been  omitted  from  this  list  for  the  reason  that  the 
formula  which  they,  by  successive  steps,  developed  is  based  on  the 
erroneous  application  to  columns  of  the  principle,  strictly  applicable 
to  beams,  that  the  greatest  possible  deflections  within  the  elastic  limit, 
of  beams  similar  as  to  section,  manner  of  loading,  and  end  conditions, 
are  proportional  to  the  squares  of  their  lengths  multiplied  by  the  elas- 


262  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Prichard.  tic  limit.  In  columns,  nnder  analogous  conditions,  the  greatest  de- 
flections within  the  elastic  limit  are  proportional  to  the  squares  of  their 
lengths,  multiplied,  not  by  the  elastic  limit,  but  by  the  differences  be- 
tween the  elastic  limit  and  the  mean  compressive  stresses  in  the  various 
columns. 

Euler's  formula  applies  only  to  long  columns,  but  he  should  be 
included  among  those  who,  by  analysis,  have  determined  facts  as 
to  short  columns,  for  the  reason  that  his  formula  carries  with  it  the 
necessary  consequence  that,  under  ideal  conditions,  columns  which  are 
too  short  to  have  Euler's  formula  apply  to  them  will  have  a  uniform 
distribution  of  stress  and  no  deflection,  up  to  the  elastic  limit,  a  con- 
dition which  is  sometimes  closely  approached  in  laboratory  tests.  In 
practice,  of  course,  the  conditions  may  be  far  from  ideal,  but  Fidler, 
Marston,  J.  B.  Johnson,  Moncrieff,  and  others  have  made  valuable  and 
instructive  analyses  of  the  efi^ect,  within  the  elastic  limit,  of  departures 
from  ideal  conditions. 

The  author  objects  to  the  practice  of  using  the  elastic  limit  as  the 
criterion  of  strength  without  regard  to  the  ultimate.  Wlien  rest  oc- 
curs between  the  periods  of  straining  beyond  the  yield  point,  the  elas- 
tic limit,  which  at  first  is  somewhat  below  the  yield  point,  can  be 
raised  somewhat  above  it,  thus  making  a  permanent  gain  in  strength, 
the  usefulness  of  which  is  greatly  lessened  by  the  fact  that  when 
structural  steel  of  the  usual  quality  is  overstrained  it  becomes  very 
4uctile. 

When  only  a  small  portion  of  a  steel  member  is  overstrained,  and 
the  conditions  are  such  that  a  very  small  flow  of  the  ductile  metal 
brings  relief,  the  overstrained  steel,  by  regaining  its  elasticity  during 
a  rest,  accommodates  itself  to  the  conditions  with  comparatively  slight 
distortion.  Thus  ductility,  combined  with  the  recuperative  powers  of 
the  steel,  may  be  useful  in  adjusting  the  length  and  shape  of  members 
and  details,  and  in  raising  the  strength  of  pins,  etc.,  but  if  the  stress 
over  the  entire  cross-section  of  a  member  is  even  slightly  greater  than 
the  yield  point,  and  there  is  no  other  direct  path  for  it  to  follow,  the 
member,  if  in  compression,  will  buckle,  unless  it  is  very  short  and  stifle, 
and,  if  in  tension,  will  elongate  so  much  that  it  will  not  only  be  ir- 
reparably injured,  but  will  cause  ruinous  distortion  in  the  remainder 
of  the  structure,  and  possibly  the  failure  of  some  adjacent  compression 
member,  to  the  supposed  weakness  of  which  the  disaster  may  be 
erroneously  attributed. 

Between  ruinous  distortion  and  collapse  there  is  a  great  difference: 
Kuinous  distortion  means  the  loss  of  the  structure,  while  collapse  may, 
in  addition,  cause  great  damage  and  loss  of  life.  The  possession  of 
strength  in  excess  of  the  yield  point,  even  though  it  be  but  temporary, 
is,  therefore,  of  some  value,  and  a  somewhat  higher  unit  stress  could 
be  allowed  in  members  which  possess  it  than  in  those  which  do  not. 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS 


263 


It  should  be  remembered,  however,  that  even  a  tension  member,  from  Mr 
some  action  which  starts  from  a  nick,  a  flaw,  a  jagged  edge,  a  thread, 
or  a  rivet  hole,  or  in  some  detail  from  no  distinctive  point,  may  break 
without  marked  elongation,  especially  under  shock. 

There  is  no  analytical  method  by  which  strength  within  the  elastic 
limit  can  be  equated  with  strength  beyond  it,  but,  other  things  being 
equal,  would  not  an  advantage  of  4  000  lb.  per  sq.  in.  fully  offset  the 
absence  of  any  strength  there  may  be  in  a  highly  ductile  metal  beyond 
that  limit?  To  assist  in  considering  this  question,  the  elongations 
beyond  the  yield  point  during  a  test  of  a  fairly  typical  eye-bar  are 
submitted  in  Table  1. 

TABLE  1. — Elongation  of  a  Typical  Steel  Eye-bar,  Measured  in  a 
Length  of  262|  In.  from  Center  to  Center  of  Pins. 


Prichard. 


Load  per 

Elongations: 

Load  per 

Elongations: 

square   inch, 
in   pounds. 

square  inch, 
in  pounds. 

Inches. 

Percentage. 

Inches. 

Percentage. 

35  000 

0.55 

0.03 

45  000 

6.92 

2.64 

36  000 

2.07 

0.79 

50  000 

9.75 

3.72 

37  000 

2.75 

1.05 

55  000 

13.60                       5.17 

38  000 

3.77 

1.44 

60  000 

30.63 

7.84 

39  000 

4.21 

1.60 

64  410 

38.25 

14.6 

40  000 

4.56 

1.74 

56  710 

40.65 

15.5 

A  strain  of  4  500  lb.  per  sq.  in.  in  excess  of  the  yield  point,  with 
a  percentage  of  elongation  equal  to  the  percentage  in  the  bar  cited,  if 
it  occurred  in  the  diagonal  eye-bars  of  a  bridge  panel  25  ft.  long  by  25 
ft.  high,  would  cause  a  distortion  of  the  bridge,  in  one  panel  length, 
of  9-2-  in.    This  would  be  a  severe  test  of  the  floor  system  and  top  chords. 

The  relation  of  the  yield  points  in  compression  to  those  in  tension 
was  well  shown  by  a  set  of  comparative  tests  by  the  late  Charles  A. 
Marshall,*  M.  Am.  Soc.  C.  E.,  a  synopsis  of  which  is  given  in  Table  2. 

In  Table  2  the  strength  of  the  steel  increases,  in  a  general  way,  as 
the  size  and  thickness  of  the  sections  are  reduced.  A  similar  variation 
in  the  strength  of  wrought  iron  was  shown,  by  tests  made  by  the  United 
States  Board  on  Testing  Iron  and  Steel,t  to  be  due  to  reduction  in 
rolling.  In  most  cases,  the  results  for  compression  are  each  an  average 
of  two  tests,  and  for  tension,  of  three  or  four  tests.  The  average  of 
the  yield  points  given  in  Table  2  for  compression  is  1  432  lb.  per  sq. 
in.  greater  than  for  the  corresponding  results  for  tension.  The  results 
of  these  tests  cannot  be  applied  directly  to  tension  in  eye-bars.  The 
fact  that  eye-bars  are  annealed  puts  them  in  a  different  class  from 
material  as  it  conies  from  the  rolls,  as  the  steel  is  softened,  some  of  the 
good   effects  of  rolling  are   taken   away,   and   the  proportion  of  yield 


*  Transactions,  Am.  Soc.  C.  E.,  VoL  XVII,  p.  i 
t  Vol.  I,  1H81,  pp.  35-45. 


204 


DISCUSSION  OK  SAFE  STRESSES  IN"  STEEL  COLUMNS      [Papers. 


Mr.  Prichard.  point  to  ultimate  is  lowered,  especially  if  the  bars  are  cooled  slowly. 
Except  in  rare  cases,  steel  as  it  comes  from  the  rolls  will  have  a  yield 
point  in  tension  exceeding  55%  of  its  ultimate  strength,  while  the 
average  of  all  the  tests  (some  570  or  more)  of  full-sized  eye-bars,  made 
during  the  last  few  years  at  the  Ambridge  plant  of  the  American 
Bridge  Company,  gives  a  yield  point  equal  to  524%  of  the  ultimate 
strength  of  the  full-sized  bar,  with  variations  above  and  below  this 
percentage.  It  is  not  wise  to  count  on  a  yield  point  of  more  than  50% 
of  the  ultimate  strength  of  the  bar. 

TABLE  2. — Comparative  Tests,  in  Tension  and  Compression. 

All  fi'om  the  same  blow  of  Bessemer  steel  as  it  came  from  the  rolls. 


Yield  Point. 

Ultimate 

strength  in 

teusion,  in 

pounds  per 

.square  inch. 

Size  and  sliape  of 
test  piece. 

Compression, 
in  pounds  per 
square  ineti. 

Tension : 

In  pounds  per 
square  inch. 

Percentage 
of  ultimate. 

4  by  ]/^  in 

Not  given. 

49  055 
Not  given. 

47  300 
Not  given. 

43  845 

46  020 
Not  given. 

42  300 

43  460 
41290 

53  800 
47  815 
47  363 
46  090 
44  417 
44  273 
44  202 
43  560 
41  527 
41447 
41415 
41060 
40  747 
40  275 
40  017 
39  397 
39  317 
39  302 
38  482 
38  310 
38  207 
38  193 
37  820 
37  580 
37  000 
36  680 
30  100 
35  917 

75.5 
68.8 
68.9 
66.8 
65.8 
64.7 
65.0 
63.6 
62.1 
61.9 
62.4 
60.4 
60.8 
60.7 
60.3 
59.1 
58.8 
59.1 
57.7 

58;2 
57.5 
57.0 

m.b 

71255 

69  390 

3bv?^in  

68  657 

%  in  round 

68  995 

3  hv  1^  in 

67  527 

68  427 

67  970 

1  in  .square 

4  by  34  in 

68  510 
66  917 

3  by  %  in 

4  by  1  in ; 

66  987 
66  230 

114  in  square 

67  973 

67  040 

114  in       "         

66  363 

14d  in       "             

42  075 
Not  given. 

42  740 
Not  given. 

39  940 

66  333 

3bv  1  in 

66  700 

66  833 

3  by  2  in        

66  537 

3  bA'  1J4  in 

66  640 

Not  given. 

2  in.  round 

38  830 

40  630 

Not  given. 

36  840 
Not  given. 

65  663 

66  400 

3  by  114  in   

66  342 

4  by  114  in 

Not  given. 
65  460 

Not  given. 

I.        n 

3  by  1^  in 

65  762 

It  appears  from  the  foregoing  that  the  higher  yield  point  in  com- 
pression, of  steel  as  it  comes  from  the  rolls,  as  compared  with  the  yield 
point  of  annealed  eye-bars,  would  about  offset  the  advantage  which  the 
latter  possesses  of  some  temporary  strength  in  excess  of  the  yield 
point,  even  when  the  same  ultimate  tensile  strength  is  specified  for 
the  eye-bars,  as  determined  by  full-sized  tests,  and  steel  for  compression 
members,  as  determined  by  specimen  tests.  A  comparison  between 
the  strength  of  steel  in  compression  and  the  tensile  strength  of  built 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS  265 

steel  members  is  a  different  matter,  and  introduces  a  somewhat  differ-  Mr.  Pilchard, 
ent  set  of  questions.    Few  engineers,  however,  unless  they  are  opposed 
to  pin  connections,  would  permit  a  higher  unit  stress  for  the  net  area 
of  a  built  tension  member  than  for  an  eye-bar. 

With  the  exception  of  columns  made  of  pipes  and  single  angles, 
practically  all  columns  are  built  of  sections  and  plates.  The  rivets 
are  usually  assumed  to  fill  the  holes  and  take  the  place,  as  far  as  com- 
pression is  concerned,  of  the  sections  they  replace.  As  a  matter  of 
fact,  they  do  not  completely  fill  the  holes,  and  it  is  very  doubtful 
whether  they  wholly  make  good  the  loss  in  section.  They  are  seldom 
placed  closer  than  an  average  of  4  in.,  and  probably  they  are  more 
than  half  as  effective  as  the  metal  they  replace.  On  this  basis,  the 
allowed  stress  per  unit  of  gross  section  of  column  area  would  be  about 
seven-eighths  of  the  allowed  stress  per  unit  of  net  section  in  tension; 
that  is,  if  16  000  lb.  per  sq.  in.  is  allowed  in  tension,  14  000  lb.  would 
be  a  corresponding  limit  for  compression.  There  are  other  considera- 
tions, however,  chief  of  which  is  the  weakening  influence  of  slender- 
ness  in  either  the  column  as  a  whole,  or  in  its  details. 

Notwithstanding  the  large  number  of  tests  that  have  been  made  in  i 

the  endeavor  to  determine  the  influence  of  slenderness  (Moncrieff,  in 
his  paper  on  "The  Practical  Column,'"'  cites  more  than  1000),*  the 
practice  of  engineers  in  this  regard,  as  shown  by  the  author,  is  very 
diverse;  from  which  it  would  appear  that  the  lessons  taught  by  the 
tests  are  not  very  definite,  or  that  they  have  not  been  generally  under- 
stood. 

A  knowledge  of  the  principles  involved  is  of  great  importance, 
both  as  a  guide  to  the  making  of  useful  tests  and  as  a  key  to  under- 
standing the  phenomena  observed.  The  theory  of  columns  has  been 
partially  developed  by  correct  analysis,  but  it  has  frequently  been 
elaborated  so  much  that  the  essential  facts  have  been  buried  under 
what  Trautwine  called  "heaps  of  mathematical  rubbish."  It  may  be 
well,  therefore,  to  present  a  concise  analysis  of  the  influence  of  length 
and  eccentricity  on  the  strength  and  stiffness  of  columns. 

Consider  a  column  with  frictionle.ss  hinged  ends,  of  length,  I,  and 
radius  of  gyration,  r,  with  constant  cross-sectional  area.  A,  subjected 
to  a  longitudinal  load,  of  intensity,  p,  acting  with  an  intentional  ec- 
centricity, e,  and  an  accidental  eccentricity,  e'. 

In  consequence  of  the  eccentricity,  there  will  be  a  primary  inten- 
tional bending  moment,  p  A  e,  o.  primary  accidental  bending  moment, 
p  A  e',  and  a  secondary  bending  moment,  p  A  ^  ;  A  being  the  de- 
flection. The  value  of  ^^  can  be  obtained  from  the  well-known  equa- 
tion: 

moment  f 

^  C  E  Ar"" ^^) 


*  Transactions,  Am   Soe.  C.  E.,  Vol.  XLV,  p.  334. 


S66  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Prichard.  in  which  E  is  the  modulus  of  elasticity  and  G  a  factor  which  varies 
with  the  shape  of  the  moment  diagram. 

The  moment  diagram  can  be  divided  into  two  parts,  the  diagram 
of  primary  moments  and  the  diagram  of  secondary  moments.  For  the 
determination  of  the  deflection  due  to  the  secondary  moments,  the 
value  of  C  will  vary  between  limits  of  which  the  upper  is  n^  and  the 
lower  depends  on  the  form  of  the  primary  moment  diagram:  if  it  is 
a  rectangle,  the  lower  limit  will  be  9.6,  while,  if  it  is  in  the  shape  of  a 
bow,  the  lower  limit  will  approach  very  close  to  the  upper  limit, 
n^  =  about  9.87.  The  limits  are  so  narrow  that  it  can  be  taken  as  n' 
without  serious  error. 

For  the  determination  of  the  deflection  directly  due  to  the  primary 
moment,  the  value  of   C  will   vary  according  to  the  conditions,  but, 

for  convenience,  it  may  be  designated  — .     (If  the  cause  of  the  primary 

z 

bending  moment  is  the  eccentric  application  of  the  load,  z  will  equal 

1.234,  and  C  will  equal  8;  but,  if  the  cause  is  a  bow-shaped  bend  in 

the  axis  of  the  column,  z  will  be  approximately  equal  to  unity.     In 

the  applications  made  subsequently  in  this  discussion,  z  is  taken  as 

equal  to  1.234,  which,  in  some  cases,  is  a  trifle  high.     The  resulting 

stresses  and  deflections,  therefore,  are  a  trifle  high,  especially  for  the 

higher  ratios  of  I  to  r.) 

Substituting  the  primary  intentional,  primary  accidental,  and  sec- 
ondary moments  in  Equation  1  gives 

zp  A  X'  (e  -\-  e')        p  A  A  X- 
A  Tt-  E  v  Atc-Ev 

To  simplify  the  development,  let 

^=  -1^ ;••; ^"^ 

This  is  Euler's  formula,  and,  as  it  facilitates  the  application  of  the 
final  equations  to  have  the  values  of  q,  which  may  be  termed  a  modulus 

of  rupture,  determined  for  various  values  of         and  tabulated.  Table  ?> 

r 

is  submitted. 

Substituting  q  for  its  value  in  Equation  2,  and  reducing,  gives 

^^^t±(i±n (4) 

q—p 
Hence  the  secondary  moment  is 

r)  A  A  =  p  A  (e  -{-  e')  — - —  z (f)) 

q—p 

Let  V  =  the  distance  from  the  neutral  axis  to  the  extreme  fiber  on 

the  concave  side  of  the  column. 

The  stress  from  bending,  in  the  extreme  fiber  on  the  concave  side 

of  the  column,  is   5 .      Hence,  if  f  =  the  combined  stress 

A  r 

in  the  extreme  fiber  on  the  concave  side. 


Papers.^      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS  267 

/  =.  p  +    (JLlZ  +  P^^    (  1  +  _JL_    ,  )     (6)  Mr.  Prtc,.a„>. 

In  investigating  physical  laws,  the  work  of  the  study  and  the 
laboratory  should  be  complementary — a  proposition  generally  conceded, 
but  practiced  too  little  in  investigating  the  mechanics  of  structures. 

TABLE  3. — Values  for  Modulus  of  Buckling.* 

7r2  E 


Values  of  the  modulus  of  buckling,  q 

E  =  29  000  000  lb.  per  sq.  in. 

I  =  length,  in  inches,     r  =  radius  of  gyration,  iu  inches. 


{'■) 


I 

r 

Values  of  q. 

■ 

I 

r 

Values  of  q. 

I 
r 

! 

Values  of  q. 

■z 

71555  000 

82 

42  567 

162 

10  906 

4 

17  889  000 

84 

40  564 

164 

10  642 

6 

7  950  500 

86 

38  700 

166 

10.387 

8 

4  472  200 

88 

36  9«0 

168 

10  141 

10 

2  862  200 

-  90 

35  336 

170 

9  904 

12 

1  987  600 

92 

33  816 

172 

9  675 

14 

1  460  300 

94 

32  393 

174 

9454 

16 

1  118  000 

96 

31  057 

176 

9  240 

18 

88:^390 

98 

29  802 

178 

9034 

20 

715  550 

100 

28  622 

180 

8  834 

22 

591  360 

102 

27  511 

182 

1      8  641 

24 

496  910 

104 

26  463 

184 

8  454 

26 

423  mo 

106 

25  473 

186 

8273 

28 

365  080 

108 

34  549     ( 

188 

8098 

30 

318  020 

110 

23  655 

190 

7  929 

32 

279  510 

112 

22  8ir 

192 

7  764 

34 

247  590 

114 

22  024 

194 

7605 

36 

220  850 

116 

21  271 

196 

7  451 

38 

198  210 

118 

20  556 

198 

1       7  301 

40 

178  890 

120 

19  876 

200 

7  155 

42 

162  260 

122 

19  230 

2(»2 

7  015 

44 

147  840 

124 

18615 

2lH 

6  878  ■ 

46 

135  260 

126 

18  029 

206 

6  745 

48 

124  230 

128 

17  469 

208 

6  616 

50 

114  490 

130 

16  936 

210 

6  490 

52 

105  850 

132 

16  427 

212 

6  368 

54 

98  155 

134 

15  WO 

214 

6  250 

56 

91  269 

136 

15  475 

216 

6  135 

58 

85  083 

138 

15  029 

218 

6  023 

60 

79  506 

140 

14  603 

220 

5  914 

62 

74  459 

142 

14  195 

222 

5  808 

64 

69  878 

144 

13  803 

224 

5  704 

66 

65  707 

146 

13  427 

226 

5604 

68 

61899 

148 

13  067 

228 

5  506 

70 

58  412 

150 

12  721 

230 

5  411 

72 

55  212 

152 

12  3&S 

232 

5  318 

74 

52  268 

154 

12  068 

234 

5  227 

76 

49  553 

156 

11  761 

236 

5139 

78 

47  045 

158 

11  465 

238 

5053 

80 

42  722 

160 

11  180 

240 

4969 

*  From  Proceedings.  Engineers'  Society  of  Western  Pennsylvania.  July.  1907,  p.  341. 


268 


DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 


Mr.  Prichard.  To  Substantiate  Equation  4,  which  differs  in  form,  but  is  actually 
similar  to  one  given  by  Moncrieff,*  a  comparison  is  submitted  between 
the  deflections  calculated  therefrom  and  the  actual  deflections  in  tests 
made  at  the  Watertown  Arsenalf  of  four  wrought-iron  columns;  two 
consisting  of  two  8-in.  channels  and  one  12-in.  cover-plate;  and  two  of 
two  10-in.  channels  and  one  13-in.  plate.  In  both  cases  the  columns 
were  latticed  on  the  side  opposite  to  the  plate,  the  length  was  74  radii 
of  gyration,  the  load  was  applied  by  three  ^-in.  pins  at  right  angles  to 
the  webs  of  the  channels,  a-nd  placed  in  the  center  of  gravity  of  the 
channels.  The  modulus  of  elasticity  was  assumed  as  27  000  000  lb.,  and 
the  value  of  q  correspondingly  determined  as  48  600  lb. 

TABLE  4. 


Test  1632: 

Test  1633: 

e-    1.62  IN. 

e=    1.64  IN. 

4  =  17.57  SQ.  IN. 

A^  17.72  SQ.  IN. 

Load, 

in 

pounds. 

Deflections,  in  inches. 

Deflections,  in  inches. 

Calculated. 

Actual. 

Calculated.              Actual. 

10  000 

0.02 

0.00 

0.02                         0.00 

20  000 

0.05 

0.01 

0.05           '              0.01 

50  000 

0.12 

0.10 

0.12                          0.09 

100  oon 

0.27 

0.22 

0.27                          0.21 

200  000 

0.61 

0.54 

0.61                            0..51 

250  0(i() 

0.83 

0.77 

0.8.3                            0.74 

280  000 

0.97 

0.95 

0.97                            0.91 

SOOOOO 

1.08 

1.15 

1.08                            1.20 

306  000 

Ult.  Load. 

307  000 

Ult.  Load. 

310  000 

1.14 

1.14                       

Test  350: 

Test 

351: 

e  =    1.7  IN. 

e  = 

.3  IN. 

A  =  12.48  SQ.  IN. 

A=  . 

Load. 

in 

pounds. 

Deflections,  in  inches. 

Deflections 

,  in  inches. 

Calculated. 

Actual. 

Calculated. 

Actual. 

10  000 

0.04 

0.0 

0.03 

0.0 

20  000 

0.07 

0.03 

0.06 

0.3 

50  000 

0.19 

0.11 

0.17 

0.12 

100  000 

0.41 

0.37 

0.38 

0.32 

150  000 

0.69 

0.67 

0.64 

0.58 

180  000 

0.89 

0.90 

0.83 

0.82 

200  000 

1.03 

1.20 

0.98 

1.05 

202  700 

Ult.  Load. 

208  200 

Ult.  Load. 

210  000 

1.11 

1.06 



*  Transactions,  Am.  Soc.  C.  E.,  Vol.  XLV,  p.  359. 

+  Reports  for  188:^,  pp.  167  and  168:  and  1884.  pp.  54  and  55. 


Papors.]      DISCUSSION  ON   SAFE   STRESSES  IN  STEEL  COLUMNS  269 

Considering  the  fact  that,  in  making  the  comparisons  in  Table  4,  Mr.  Prichard. 
no  allowance  was  made  for  unintentional  eccentricity  or  pin  friction, 
the  agreement  is  as  close  as  could  reasonably  be  expected.  Monerieff, 
in  connection  with  his  equation  for  deflection,  previously  referred  to, 
gives  a  large  number  of  comparisons  between  deflections  obtained  by 
applying  his  equation,  and  those  observed  in  tests,  which  tend  strongly 
to  establish  its  substantial  accuracy. 

The  maximum  compression  in  the  extreme  fiber  of  the  column 
bearing  the  test  number  1632,  has  been  calculated  for  the  ultimate  load 
by  Eqiuition  6,  which  gives  36  600  lb.  per  sq.  in. 

The  amoimt  of  the  unintentional  eccentricity  will  fluctuate  greatly 
in  practice,  but,  in  devising  rational  formulas  for  use  in  designing, 
either  by  direct  application  or  through  empirical  formulas  founded  on 
them,  the  greatest  amount  which  is  reasonably  possible  with  ordinary 
care  should  be  assumed.  The  amount  assumed  should  cover  inac- 
curacies in  boring  pin  holes,  the  shift  in  the  position  of  the  axis  from 
over-runs,  and  shortages  in  sectional  area  as  compared  with  the  area 
of  the  sizes  specified,  inequalities  in  the  modulus  of  elasticity  in  dif- 
ferent parts  of  the  cross-section,  curves  or  kinks  in  the  axis,  and 
potential  curves  or  kinks  in  the  axis  from  the  relief  of  internal  stresses. 
Owing  to  internal  stresses  produced  by  cold-straightening  or  otherwise, 
the  metal  is  likely  to  be  overstrained  in  spots  before  that  in  the 
main  body  of  the  column  reaches  the  elastic  limit.  The  internal 
stresses  may  be  relieved  to  some  extent  by  overstraining  followed  by  a 
rest,  but  the  column  is  likely  to  have  a  permanent  deflection  as  a  re- 
sult thereof.  From  some  causes,  such  as  inaccuracies  in  pin  holes, 
short  columns  are  likely  to  have  as  much  accidental  eccentricity  as 
long  ones;  while,  from  other  causes,  such  as  initial  curvature  of  the 
axis,  the  probable  limit  of  eccentricity  will  vary  with  the  length.     The 

e'  V 
following  value  for  the  factor,  -^-,  in  Equation  0  is  suggested: 

r' 

e^_  1       _Z 
T^  "~  10  "*~  700  7 
For   ordinary   built   coliimns   with   pin   connections,    in   which   the 
relation  of  V  to  r  is  about  as  7  is  to  5,  the  eccentricity  corresponding 
to  Equation  7  is : 

e'  =  0.07  r  -f  0.01  I (,S) 

For  a  column  with  a  radius  of  gyration  of  5  and  a  length  of  500 
in.,  the  eccentricity  given  by  Equation  8  is  about  |  in. 

From  Equations  6  and  ?,  a  formula  can  be  deduced  which  will  give 
the  load,  p,  per  unit  of  column  area,  but  such  a  formula  is  not  sub- 
mitted, for  the  reasons  that  it  is  so  complicated  and  difficult  of  ap- 
plication that  it  is  of  no  practical  value,  and  the  results  which  it  gives 
for  columns  having  a  length  of  less  than  100  radii  of  gyration  can  be 


2   -  T7T  +  ^7^ (') 


270 


DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 


Mr.  Prichard.  jipproximated  closely  by  short  empirical  methods.     The  method  sug- 
gested is  as  follows: 

For  structural-steel  columns  with  hinged  ends,  the  stress  per  square 
inch,  in  pounds,  in  the  most  compressed  fiber,  from  combined  direct 
compression  and  intentional  primary  bending  moment,  shall  not  exceed 

f 
35  000  —  1.5   2 

1- (!») 

Factor  required. 

For  a  factor  of  2.5,  the  expression  becomes 

r- 


14  000  —  0.0 


(1(1) 


From  Equations  G  and  7  it  is  evident  that  the  minimum  value  which 
can  be  assigned  for  the  stress  from  the  accidental  bending  moment  is 
O.lp^  from  which  it  follows  that  a  limitation  of  35  000  lb.  per  sq.  in. 
for  combined  direct  compression  and  intentional  primary  bending 
moment  corresponds  to  a  limitation  of  38  500  lb.  per  sq.  in.  for  stresses 
from  all  sources.  Hence,  if  proportioning  by  Equation  9,  with  a  factor 
of  one  for  maximum  possible  loads,  gives  results  closely  in  accord  with 
theory,  as  claimed,  the  stress  per  square  inch  in  columns  thus  pro- 
l)ortioned,  as  determined  by  Equations  6  and  7,  should  be  close  to 
38  500  lb.     How  close  they  come  to  this  amount  is  shown  by  Table  5. 

TABLE  5. — Maximum  Stresses  per  Square  Inch,  Determined  by 
Theory  (Equations  6  and  7)  in  Columns  Proportioned  by  Rule 
(Equation  9,  with  Factor  of  One),  for  Various  Lengths  and 
Intentional  Eccentricities. 


The  intentional  primary  bending  stresses,  in  terms  of  the 

LOAD,   are   given  at  THE  HEAD   OF   EACH  COLUMN. 

I 

r 

0 

O.lp. 

0.5p. 

P- 

Pounds. 

Pounds. 

Pounds. 

Pounds. 

0 
25 
50 
75 
100 

38  500 
39100 

39  100 
39  400 

38  700 

38  200 

39  000 
39  200 
39  800 
38  800 

37  300 

38  100 
38  600 
38  500 
33  900 

36  750 

37  300 
.37  500 
36  300 
30  700 

The  agreement  between  theory  and  rule,  indicated  by  Table  5,  is 
close,  except  for  long  columns  and  great  eccentricities,  for  which  the 
rule  requires  a  heavier  column  than  theory. 

Some  opportunity  for  comparison  between  the  theory  and  assump- 
tions outlined  on  the  one  hand  and  experiments  on  the  other  is  afi^orded 
l)y  the  tests  of  mild  steel  columns  with  pivoted  ends  made  by  Professor 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS 


271 


Tetmajer."'  Of  the  columns  tested,  the  lengths  of  27  did  not  exceed  Mr.  Pnchard. 
100  radii,  and  they  were  loaded  without  intentional  eccentricity.  The 
ultimate  load  in  all  cases  was  greater  than  indicated  by  Equation  9, 
and  less  than  required  under  ideal  conditions  for  an  elastic  limit  of 
41  000  lb.  and  a  modulus  of  elasticity  of  29  000  000  lb.  Table  6  is  a 
comparison  of  the  ultimate  loads  given  by  Equation  9  and  the  lowest 
of  the  ultimate  loads  shown  by  the  tests. 


TABLE  6. 


I 

r 

Equation  9. 
Pounds. 

Tests. 
Pounds. 

I 
r 

Equation  9. 
Pounds. 

Tests. 
Pounds. 

80 
40 
50 

33  650 
33  600 
31  250 

39  000  about. 

38  000       " 
32  000       " 

70 
75 
92 

37  600 
26  560 
23  300 

39  000  about. 

28  000 

33  000       '• 

Eor  columns  of  greater  length  than  100  radii  of  gyration,  stiffness 
rather  than  strength  is  the  governing  consideration.  For  this  reason, 
the  loads  allowed  by  Equation  9  are  preferable  to  those  allowed  by  the 
theory  of  column  strength. 

To  show  the  relative  stiffness  of  columns  100  radii  and  longer  in 
length,  when  proportioned  by  Equation  9  with  a  factor  of  one,  the  de- 
flections have  been  determined  by  Equations  4  and  8  for  columns  with- 
out intentional  eccentricity,  with  a  radius  of  gyration  of  one,  and 
various  lengths  as  shown  in  Table  7. 


TABLE  7. 

_ 

Length, 

in 
inches. 

Deflection, 

in 

inches. 

Ratio 

of  deflection 

to  length. 

Length, 

in 
inches. 

Deflection, 

in 

inches. 

Ratio 

of  deflection 

to  length. 

100 
110 

0.486 
0.550 

1:206 
1:200 

i 

120 
130 

0.485 
0.327 

1:247 
1:400 

It  will  be  noticed  that  the  loads  allowed  by  Equation  10  for  columns 
up  to  a  length  of  100  radii  of  gyration  are  about  one-sixth  greater  than 
those  allowed  by  the  author,  but  that  there  is  a  radical  difference  for 
longer  columns.  The  objection  among  engineers  to  columns  longer 
than  100  radii  is  largely  sentimental.  For  the  same  load,  a  column 
with  a  length  of  120,  130,  or  140  radii,  proportioned  by  Equation  10, 
in  consequence  of  its  greater  sectional  area,  is  stiffer  and  stronger  than 
one  of  a  length  of  100  radii. 

*  Transactions,  Am.  Soc.  C.  E.,  Vol.  XLV,  p.  404. 


272  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Prichard.  For  derricks,  poles,  and  other  equipment  used  in  building  and  erect- 
ing, much  longer  struts  are  used  than  in  bridge  work,  and,  when  the 
loads  are  kept  within  rational  limits,  the  flexibility  of  the  struts,  by 
permitting  them  to  absorb  impact,  is  an  element  of  safety.  In  the 
ordinary  affairs  of  life,  long  struts  are  used  as  a  matter  of  course. 
The  engineer  who  becomes  alarmed  at  long  struts  in  structures  will 
bear  his  whole  weight  on  a  walking-stick,  many  times  as  flexible  as 
steel,  with  a  ratio  of  Z  to  r  of  200.  Ample  provision  should  be  made  in 
horizontal  and  inclined  struts  for  the  stresses  from  their  own  weight. 
The  frequent  neglect  to  make  such  provision  in  long  struts  has  doubt- 
less had  something  to  do  with  the  prejudice  against  them. 

As  regards  columns,  the  greatest  need  for  caution  to-day  is  in  pro- 
portioning short  stiff  ones,  which,  to  an  engineering  public  accus- 
tomed to  gauge  permissible  unit  stress  by  the  ratio  of  length  to  radius 
of  gyration,  have  an  appearance  of  strength  not  borne  out  by  their 
details,  and,  if  their  ends  are  square  or  fixed,  they  are  subject  to 
strains  from  imperfect  butt  joints,  or  to  secondary  stresses  produced 
by  the  deformation  of  connecting  floor-beams,  etc.  Such  stresses  are 
greater  for  short  than  for  long  columns,  on  account  of  their  greater 
stiffness.  In  consequence  of  these  facts,  it  is  suggested  that  the 
average  load  from  direct  compression  per  square  inch  of  cross-sectional 
area  should  not  exceed  13  000  lb. 

With  the  double  requirement  of  Equation  10  and  a  13  000-lb.  limi- 
tation for  direct  compression,  if  the  permissible  loads  are  plotted  to  a 
scale  for  various  ratios  of  I  to  r,  the  line  indicating  the  maximum 
permissible  loads  will  suddenly  change  its  direction  at  a  length  some- 
what less  than  41r,  depending  on  the  amount  of  the  primary  bending 
moment.  This  is  a  feature  which  the  author  seems  to  consider  objec- 
tionable. It  is  entirely  natural,  however.  Radically  different  condi- 
tions govern  the  strength  of  very  short  and  very  long  columns,  and 
the  loci  representing  the  loads  under  these  radically  different  condi- 
tions will  intersect  sharply.  If  there  is  no  intentional  primary  bending 
moment  from  eccentricity  or  transverse  loading,  13  000  lb.  per  sq.  in., 
unreduced,  will  govern  for  columns  of  shorter  length  than  41r,  and 
Equation  10  for  columns  of  greater  length,  but  if  there  is  an  inten- 
tional bending  moment,  both  requirements  should  be  applied  to  de- 
termine the  governing  one. 

One  of  the  assumptions  from  which  Equation  10  was  developed 
was  that  of  frictionless  hinged  ends.  When  there  is  no  primary  bend- 
ing moment,  any  friction  on  the  pins,  according  to  strict  theory,  will 
fix  the  ends;  hence,  it  is  not  surprising  that  friction  in  pins  is  very 
potent  in  increasing  the  resistance  of  columns  to  direct  load  in  care- 
fully devised  and  conducted  tests.  Such  friction,  however,  is  a  very 
poor  reliance  in  practice,  as  it  may  be  overcome  by  a  little  eccentricity 
or  shock. 


Pai)ers.]      DISCUSSION  ON   SAFE  STRESSES  IN  STEEL  COLUMNS  273 

Under  ideal  conditions,  a  column  with  strictly  fixed  ends  has  tlie  Mr.  Prichanl 
same  strength  as  a  column  of  half  its  length  with  the  same  cross- 
section  and  pivoted  ends.  In  practice,  however,  the  assumption  of  fixed 
ends  is  never  wholly  realized.  The  appearance  of  having  the  ends 
fixed  is  frequently  deceptive,  as  compression  members  on  opposite 
sides  of  a  joint  may  deflect  in  opposite  directions  in  such  a  way  that 
the  point  of  contrary  flexure  comes  very  near  to  the  center  of  the  joint, 
which  condition  is  equivalent  to  pivoted  ends.  For  this  reason,  no 
easement  in  the  reduction  for  length,  as  given  in  Equation  10,  is 
recommended  in  designing  new  columns  with  seemingly  fixed  ends. 
In  determining  the  safe  strength  of  a  column  in  an  existing  structure, 
however,  if  it  is  evident  that  the  ends  are  well  fixed,  it  might  be  as- 
sumed that  the  column  is  as  stiff  as  it  would  be  if  it  were  about  three- 
fourths  as  long  and  had  frictionless  hinged  ends.  For  such  an  assump- 
tion, Equation  9  would  become : 

Allowed  compression  per  square  inch  in  any  fiber  from  comliined 
direct  compression  and  intentional  primary  bending  moment  for 
columns  with  ends  fixed  equals 

36  000  —  0.85  -, 


Factor  required. 

In  comparing  this  discussion  with  the  writer's  paper,*  entitled  "The 
Proportioning  of  Steel  Railway  Bridge  Members,"  it  will  be  noticed 
that  the  greatest  compression  now  recommended  is  one-fifteenth  less 
than  in  the  paper  referred  to,  in  addition  to  which  a  limitation  for 
direct  compression  to  13  000  lb.  per  sq.  in.  is  now  recommended.  This 
change  is  the  result  of  a  further  consideration  of  the  subject,  in  the 
light  of  the  Quebec  Bridge  disaster  and  the  general  discussion  regard- 
ing columns  which  followed  it,  including  the  paper  by  Mr.  Worcester. 
It  should  be  stated,  however,  that  it  has  always  been  the  writer's  prac- 
tice in  designing  columns  to  give  close  attention  to  the  details  and  to 
the  make-up  of  the  section,  and  to  make  a  liberal  reduction  in  the 
allowed  unit  stresses  when  the  unsupported  width  of  a  plate  exceeded 
32  times  its  thickness.  The  following  requirement  as  to  the  unsup- 
ported width  of  plates  is  siiggested: 

If  the  unsupported  width,  w,  of  any  plate  in  a  column  is  more 
than  32  times  its  thickness,  i,  the  permissible  stress,  as  given  by  Equa- 
tion 10,  shall  be  reduced  by  multiplying  it  by  the  following  expression: 

2  ?o2 
16  000  —  -^ 

Permissible  stress,  as  given  by  Equation  10,    X  

14  000 

Horace  E.  Horton,  M.  Am.  Soc.  C.  E.  (by  letter).— Mr.  Worcester's  Mr.  Horton. 

compilation   of  results   of  tests  on   full-sized  wrought-metal  compres- 

*Proceedings,  Enginews  Society  of  Western  Pennsylvania,  July,  1907. 


274  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Horton.  sion  members  is  very  interesting  and  instructive,  and  is  opportune. 
There  is  an  awakened  interest  in  the  subject  at  this  time. 

Wliile  the  writer  approves  unhesitatingly  Mr.  Worcester's  criticism 
on  using  excessive  unit  stress  on  members  with  short  radii  lengths, 
he  knows  no  physical  reason  for  limiting  compression  members  to  a 
length  of  100  radii. 

Mr.  Worcester  has  chosen  to  make  his  platting  of  tests  for  steel  on 
the  basis  of  four-fifteenths  of  the  ultimate  strength.  For  obvious  rea- 
sons, the  writer  uses  the  same.  Mr.  Worcester  has  used  12  000  lb.  per 
sq.  in.  as  his  unit  value  in  compression,  and  the  writer  naturally  uses 
the  same  stress,  with  the  reservation  that  the  unit  stress  (tension)  is 
IJ  X  the  compression,  that  is,  16  000,  in  this  case. 

The  diagram,  Fig.  2,  gives  all  the  tests  of  steel  members  shown  by 
Mr.  Worcester,  also  tests  of  six  members*  by  J.  A.  L.  Waddell,  M. 
Am.  Soc.  C.  E.,  and  seven  testsf  by  Mr.  C.  P-  Buchanan,  and,  further, 
six  tests  by  the  Chicago  Bridge  and  Iron  Works,  on  8  by  8  by  x^H"i"- 
angles. 

On  this  diagram  a  straight  line  is  drawn   through   the  center  of 

gravity  of  the  group  of  tests,  and  is  expressed  by  11300  —  35  ^_  ,  also 

the  formula  for  loading,  as  indicated  by  C.  L.  Strobel,  M.  Am.  Soc. 
C.  E.,  in  his  paper,  "Experiments  Upon  Z  -Iron  Columns,":}:  wherein 
was  first  laid  down  the  necessity  of  "sawing  off"  the  unit  stress  for 
short  radii  length,  in  this  case  to  8  000  lb.  per  sq.  in.,  and  also  the  first 

appearance  of  the  straight-line  formula,  which  was  10  600  - —  30     . 

The  writer  has  platted  Mr.  Worcester's  formula,  based  on  12  000, 
and,  as  a  protest  against  the  attempted  "amputation"  for  radii  lengths 
of  more  than  100,  there  is  also  platted  the  Hodgkinson-Gordon-Rankine 
formula,  as  given  by  Mr.  Worcester,  based  on  16  000,  with  a  lower  di- 
visor of  8  000.  It  will  be  noticed  that  the  platted  lines  come  tangent 
at  85  radii,  and  the  two  curves  come  somewhat  above  the  center  of 
gravity  of  the  tests.  However,  as  these  values  are  high  for  short  radii 
lengths,  clearly  indicating  the  necessity  of  "sawing  off,"  the  writer 
offers  the  Hodgkinson-Gordon-Rankine  formula,  based  on  12  000,  with 
a  lower  divisor  of  12  000.  When  "sawed  off"  at  two-thirds  of  the  unit 
stress  (that  is,  10  666  lb.  per  sq.  in.  as  the  ultimate  value  in  com- 
pression), and  intersecting  the  curve  at  40  radii  length,  it  looks  both 
sane  and  safe. 

The  curve  produced  as  platted,  clearly  indicates  what  is  intended, 
a  value  well  below  the  average  of  the  tests,  and  the  value  reduced  so 
that  one  may  believe  it  to  be  reliable  as  the  radii  lengths  are  extended, 
even  to  200. 

*The  record  of  these  tests  appeared  in  Engineering  Nen-s,  January  16th,  1008. 
^Engineering  Newft,  December  20th,  1907. 
XTrnnsaclionx,  Am.  Soc.  C.  E.,  Vol.  XVIII.  p.  103. 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS 


275 


Mr.  Hortoii. 


27G 


DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 


Mr.  Horton.  The  limitation  to  100  radii  length  of  compression  members  can  only 
be  urged  because  of  our  want  of  knowledge,  but  practice  and  experi- 
ence show  that  the  greatest  hazards  are  with  short  radii  lengths,  due 
to  want  of  proper  proportion  and  cohesion  of  parts,  and  the  tendency 
toward  using  material  which  is  too  thin. 

Unfortunately,  engineering  discussion  as  to  the  unit  value  of 
compression  members  has  been  almost  entirely  on  formulas,  and  not 
on  the  physical  column. 

As  to  the  physical  compressive  member,  Mr.  Buchanan  gives  a  re- 
port, with  full  details  of  tests  to  destruction,  of  nineteen  full-sized 
bridge  members  as  built  for  actual  use  in  structures — twelve  of  iron 
and  seven  of  steel.  The  first  noticeable  thing  is  that  radii  length  has 
no  significance,  in  fact,  members  having  a  length  of  83  radii  were  as 
strong  as  any  tested,  and  much  stronger  than  many  of  less  than  40 
radii,  and  members  of  97  and  even  120  radii  were  a  good  average  in 
the  whole  group  of  tests. 

The  average  ultimate  strength  of  seven  steel  columns  is  31  900  lb. 
per  sq.  in.  The  average  crippling  strength  is  23  800  lb.  per  sq.  in. ; 
the  average  elastic  limit  is  19  700  lb.  per  sq.  in. 

The  ultimate  strength,  crippling  strength,  and  elastic  limit,  in  the 
foregoing  tests  as  reported,  indicate  a  value  of  scarcely  more  than 
50%  of  the  value  of  the  steel  in  tension.  This  is  startling,  with  our 
knowledge  of  specifications  permitting  the  use  of  steel  in  short  radii 
lengths  for  approximately  the  same  stress  as  in  tension.  Mr. 
Buchanan's  tests  are  given  in  Table  8. 

TABLE  8. 


Averaee  of 
Buchanan's  tests. 

I 
r 

Average  of 

T.  H.  and  J.  B.  Johnson's 

formula. 

Crippling  load. 

Actual 

crippling 

load. 

Below 
estimate. 

4  tests,  'Z,-ha,r  columns 

96 
43.6 

28  537 
33  125 

21  700 
20  730 

6  837 

15  tests,  trough  and  channel  col- 
umns    

12  395 

The  four  Z-bar  columns  lack,  on  an  average  from  the  computed 
crippling  load,  essentially  half  as  much  as  did  the  fifteen  trough  and 
channel  sections  (the  last  with  less  than  half  the  radii  length),  and 
yet  they  actually  stood  a  greater  load. 

From  the  photographs  of  the  members  taken  after  the  tests,  it  is 
seen  that  four  Z-bar  column  struts  yielded  as  a  whole  by  flexure.  The 
fifteen  other  members  yielded  by  some  order  of  wrinkling  or  failure 
in  individual  parts.  It  is  further  noticeable  that  the  sections  of  the 
Z  -bar  columns  were  much  thicker,  relatively.  It  is  clearly  apparent 
that  compressive  members  which  fail  by  wrinkling  fail  at  less  load  per 
unit  than  those  which  fail  by  flexure. 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS  277 

In  the  photographs  of  the  material,  after  the  tests  by  Mr.  Waddell  Mr.  Horton. 
referred  to  above  (all  of  the  same  cross-section),  it  is  noticeable  that 
members  81  radii  long  failed  by  flexure  while  those  27  radii  long  failed 
by  wrinkling.  From  the  Buchanan  tests  there  is  abundant  evidence 
to  conclude  that  the  best  results  are  obtained  when  the  member  yields 
by  flexure.  From  Mr.  Waddell's  tests  there  is  evidence  that,  for  the 
best  results  at  27  radii,  the  material  must  needs  be  thicker  than  for 
81  radii.    Here  is  a  suggestion,  to  be  enlarged  on  later. 

Wliile  the  radius  of  gyration  has  use,  as  indicating  the  value  of  a 
strut,  there  is  much  to  show  that  there  are  many  other  conditions  of  as 
great  importance  as  the  radii  length. 

The  radius  of  gyration  will  modify  and  hold  in  check  any  disposi- 
tion to  use  material  of  undue  thickness,  but  the  radius  of  gyration  has 
to  be  held  in  check  unless  too  thin  material  be  used.  The  radius  of 
gyration  of  a  transverse  element  of  the  column  may  be  used  as  such  a 
check. 

The  composite  nature  of  the  compression  member  directly  reduces 
its  unit  value,  as  compared  with  tension,  a  very  material  amount.  This 
is  directly  traceable  to  the  possible  rivet  efficiency  connecting  parts, 
and  it  is  undoubtedly  a  fact  that  rivets  are  driven  much  too  far  apart. 
Two  or  three  times  as  many  rivets  would  surely  give  better  results. 
Rivet  connections  between  multiple  plates,  or  plates  and  angles,  form- 
ing a  compression  member,  to  reduce  the  tendency  to  wrinkling,  are 
clearly  different  from  tension  connections,  and  the  efficiency  has  to 
be  considered  locally. 

With  material  half  as  thick  as  the  rivet  diameter,  an  efficiency  of 
50%  may  be  obtained  by  pitching  the  rivets  at  2.3  diameters;  but, 
with  material  the  thickness  of  the  rivet  diameter,  and  rivets  pitched 
at  2  diameters,  an  efficiency  of  connection  of  30%  is  all  that  is  possi- 
ble. In  practice,  rivets  are  generally  driven  with  three  times  as  much 
pitch  as  here  indicated,  and  the  assertion  may  be  made  that  the  effi- 
ciency of  the  connection  of  parts  by  rivets  scarcely  exceeds  12  per  cent. 

The  cross-section  of  the  compression  member  is  unquestionably  of 
great  significance.  The  proportions  of  the  material  in  the  flanges  and 
its  width  and  thickness  undoubtedly  have  paramount  importance. 

The  compression  member,  with  ever-increasing  tendency  in  the  evo- 
lution of  design,  has  developed  with  one  or  more  open  sides  on  which 
lattice  bars  are  used. 

The  proportions  of  such  lattice  bars,  their  connections  to  the 
columns,  and  their  relation  to  a  force  acting  through  the  compression 
member  form  a  very  material  and  important  element,  second  to  none 
in  the  design.  At  the  present  time,  there  are  in  the  technical  press 
many  letters  from  correspondents,  with  elaborate  formulas  in  which 
E  represents  the  modulus  of  elasticity,  and  e  the  eccentricity.  As  e, 
eccentricity,  is  arbitrarily  assumed,  the  writer  prefers  to  assume  a  per- 
centage of  the  compression  through  the  column,  and  call  it  shear. 


278  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Horton.  The  difference  between  the  eccentricity  discussed  and  the  shear 
outlined  is  as  follows:  Eccentricity  is  an  assumption  without  refer- 
ence to  the  magnitude  or  amount  of  the  force  acting  on  the  member, 
while  the  shear  is  a  direct  percentage  of  the  force  acting  on  the  mem- 
ber. One  leads  to  the  discussion  of  how  accurate  the  workmanship  of 
the  column  may  be,  or  is.  The  other  asserts  the  fact  that  there  must 
be  some  relation  between  the  force  acting  through  a  compression  mem- 
ber and  its  disposition  to  "side-step."  This  uncertainty  is  not  caused 
by  faulty  workmanship,  but  comes  from  a  want  of  research  and  knowl- 
edge. 

In  all  the  years  past  the  whole  discussion  and  the  specifications  for 
compression  members  have  absolutely  ignored  both  shear  and  eccen- 
tricity as  items  to  consider,  except  in  what  has  appeared  within  a  very 
short  period,  and  there  is  no  evidence  that  our  workmanship  has  espe- 
cially deteriorated  in  the  immediate  past,  but  there  is  reason  to  hope 
that  our  knowledge  of  design  may  be  enlarged. 

Figs.  3  to  15  are  given  in  order  to  indicate  to  the  eye  the  relation 
of  various  sections  expressed  by  the  radius  of  gyration;  each  section 
has  the  same  cross-section,  namely,  12  sq.  in. 

Fig.  3  is  a  solid,  3.46  in.  on  a  side,  radius  of  gyration  =  1. 
Fig.  4  is  a  hollow  square,  5  in.  on  a  side,  metal  f   in.  thick, 

radius  of  gyration  =^  2. 
Fig.  5  is  a  hollow  square,  7|  in.  on  a  side,  metal  if  in.  thick, 

radius  of  gyration  =  3. 
Fig.  6  is  a  hollow  square,  10^-  in.  on  a  side,  metal  ^f  in.  thick, 

radius  of  gyration  =  4. 
Fig.  7  is  a  hollow  square,  12|  in.  on  a  side,  metal  ^f  in.  thick, 
radius  of  gyration  =  5. 
There  are  changes  in  the  radius  of  gyration  of  from  1  to  5,  with 
the  same  cross-section,  with  a  diminishing  thickness  of  the  material, 
and  an  increasing  unit  value  of  the    material    by    all    compression 
formulas. 

Figs.  8  to  15,  inclusive,  are  interesting  as  indicating  12  sq.  in.  of 
section,  in  quite  familiar  shapes,  with  a  radius  of  gyration  of  6. 

Appended  to  compression  formulas  it  is  quite  usual  to  find  a  limi- 
tation of  thickness  to  width,  of  1  to  30,  and  Figs.  8  to  15  can  only  be 
objected  to  on  this  limitation,  and  not  as  to  the  radius  of  gyration. 

According  to  Mr.  Worcester's  curve,  as  platted  for  working  loads 
on  columns  10  ft.  long,  Figs.  Y  to  15,  inclusive,  may  be  worked  for 
11  800  lb.  per  sq.  in. ;  Fig.  6  for  11  700  lb. ;  Fig.  5  for  11  300  lb. ;  Fig.  4 
for  10  400  lb.  per  sq.  in.;  and  Fig.  3  for  zero.  With  this  conclusion 
the  writer  does  not  agree.  Figs.  4  or  5,  undoubtedly,  will  carry  the 
largest  load  of  any  of  the  sections,  3  to  15,  inclusive,  at  10  ft.  long, 
while  Fig.  3  will  undoubtedly  be  a  close  second. 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS 


279 


Mr.  Horton. 


-15)4- 


rr     (Tj 


;/-- 


S..B  a 


fc 


\ 


V25K- 


3: 


H  p  I' 


ri/- 


-2.53^- 


H 
■11— 


J 


cc 

'■ 

p 

^ 

^ 

s 

a 

i< 

tg 

-i 

c 

<r.^' 

^-H 

L-4^ 1 

Ul 

•^ 

t 

a 

* 

^ 

03 

■^+° 

,, 

I 

9  E 


•2  ^i: 


3 

M® 

4^ 

11 

rf  IK 


B 


280  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Horton.  The  Writer  would  extend  the  radius  of  gyration  to  the  elements 
making  up  the  cross-section  of  a  column,  thereby  limiting  the  thick- 
ness of  the  material.  The  radii  length  on  a  transverse  section  of  a 
plate  should  never  be  more  than  the  radii  length  of  the  column  of 
v;hich  the  plate  is  a  part,  or  if  it  is,  such  parts  should  be  used  at  a 

decreased  unit  stress,  found  by  substituting  the  value  of      thus  obtained 

in  the  general  formula. 

Angles  should  be  one-fifth  the  size  of  the  transverse  dimensions  of 
a  member,  and  not  less  than  the  thickness  of  the  plates. 

In  the  case  of  columns   with  projecting  portions,  such   as  angles, 

Z's,  etc.,        (where  I  =  projection  and  r  =  radius  corresponding  to 
r 

thickness)  must  be  doubled  and  substituted  in  formulas. 

Where  a  part  is  made  of  several  thicknesses  riveted  together,  the 
transverse  radius  of  such  a  part  will  be  taken  as  the  radius  of  the 
same  as  though  solid  and  divided  by  the  square  root  of  the  number 
of  pieces  used. 

The  radii  length  of  a  lattice  panel  or  the  pitch  of  the  lattice,  with 
the  radius  of  gyration  of  neutral  axis  parallel  with  central  line  of  the 
web,  of  a  built  channel  or  similar  section,  should  never  exceed  the 
radii  length  of  the  entire  member.  The  lattice  need  not  exceed  two 
diameters  of  the  rivet.  The  radii  length  of  the  lattice  between  the 
connections  should  not  exceed  the  radii  length  of  the  member  on  which 
the  lattice  is  used. 

The  lattice  should  have  the  ability  to  carry  shear,  assuming  the 
column  to  be  supported  at  its  two  ends  or  in  the  center : 

1. — At  the  unit  strains  allowed  in  the  column  itself,  an  assumed 
uniform  load  equal  to  10%   of  the  load  sustained  by  the 
column ; 
2. — At  a  unit  stress  of  half  the  above,  the  weight  of  the  column 
itself. 

The  writer  wishes  at  this  time  to  emphasize  his  faith  and  belief  in 
proportion — the  "Rule  of  Three"  of  our  ancestors.  It  is  the  funda- 
mental basis  of  comparison  in  all  things. 

Table  9  is  an  outline  for  five  2-built  channel  lattice  columns. 

Each  column  is  in  exact  proportion,  by  the  ratio  of  2,  in  all  its 
three  (and  more)  dimensions,  to  the  next  of  the  series.  It  follows  at 
once  that  the  cross-section  of  the  columns  will  be  as  the  square,  and, 
for  the  same  radii  length,  their  weight  as  the  cube.  The  writer  has 
outlined  for  the  center  of  this  group  of  five  columns  a  rationally  pro- 
portioned 12-in.  2-built  channel  column  having  a  section  of  23  sq.  in.; 
he  has  also  doubled  it,  and  doubled  it  again.  He  has  also  divided  the 
12-in.  2-built  channel  column  in  each  of  its  dimensions  by  2  and  by  4, 
and  in  this  tabulation  by  direct  proportion  there  are  five  2-built  chan- 


Papers.]      DISCUSSION  ON  SAFE  STllESSES  IN  STEEL  COLUMNS 


281 


nol  columns.     There  is  every  reason  to  believe  that  the  3-in.   2-built  Mr  Horton 
1 


channel  column,  at 


4  096 


part  of  the  weight  for  a  proportional  radii 


length  of  the  48-in.  2-built  channel  column,  can  be  investigated  with 
reasonable  certainty  as  to  any  in  the  group  of  columns  that  are  in  di- 
rect proportion  in  all  their  elements,  that  is,  size  of  rivets,  size  of 
lattice,  and  pitch  of  rivets;  and  it  is  in  this  way  that  research  can  be 
carried  out  at  comparatively  trifling  cost.  With  the  testing  machines 
already  available,  the  truth  can  be  developed  as  to  any  or  all  of  the 
moot  questions  as  to  value  of  cross-sections  and  radii  length. 

TABLE  9. 


Built 
channel 
columns. 

Section. 

Area,  in  square 
inches. 

m 

3 

ai 

> 

(5 

* 

1 

0) 

3 

f  R.  of  gyration, 
^"    neutral  axis, 
I  parallel  to  center 
^      line  ot  web. 

"Is  ^ 

Plates. 

Angles. 

3-in. 

6-in. 
12-in. 
24-in. 
48-in. 

2  -    3  X  i 
2  -    6  X  J 
2  -  12  X  i 
2  -  24  X  1 
2  -  48  X  2 

4—     IX      f  X  g 
4—    11  X    1^  X  J 
4-3    X    3    X  ^ 
4-6    X    6    XI 
4-12    X  12    X  2 

1.44 

5.75 
23 
92 
368 

1.08 
2.17 
4.33 
8.66 
17.32 

sin. 

I-in. 
li-in. 
3  -in. 

f  Xi 

Hx^ 

3    X  1 

6    X2 

0.22 
0.43 
0.86 
1.73 
3.55 

9  ft. 

18  ft. 

36  ft. 

72  ft. 

144  ft. 

Table  10  is  a  second  compilation  for  five  columns  having  the  same 
areas  and  dimensions  as  the  columns  in  Table  9. 

TABLE  10. 


Section. 

3    . 

CTCO 

te  a) 

< 

i 

CO 

3 

•3 

CO 

Rivets. 
Double  lattice. 

R.  of  gyration, 

neutral  axis, 

parallel  to  center 

line  of  web. 

Built 

channel 

columns. 

Plates. 

Angles. 

c 

48in. 
24-in. 
12  in. 

6-in. 

3-in. 

8  -  48  X  1 
8-24  X  in 

8—12  X;/j 

8-  exs'i 

8-  3XtIb 

4  -  4J  X  4i  X  1 
4  —  2i  X  2i  X  ^ 
4-11  XUXi 
4  -  t'b  X  i»B  X  J 

4  -  s'l  X  5*5  X   i'b 

368 

92 

23 
5.75 
1.44 

14.82 
7.41 
3.71 
1.85 
0.93 

1-in.      4  X  i 
A-in.      2x8 
i-in.     1  X  ?e 
i  in       i  X  ii 
iVin-     i  X  b\ 

1.38 
0.69 
0.35 
0.17 
0.09 

124  tt. 
62  ft. 
31  ft. 
16  ft. 
8  ft. 

The  columns  in  this  group  will  not  require  a  testing  machine,  be- 
cause when  we  have  divided  down  from  48-in.  2-built  channel  columns 
to  the  12-in.  2-built  channel  columns,  23  sq.  in.  in  area,  and  find  four 
12-in.  plates  massed  together  making  12  by  |  in.  of  metal  combined 


282  DISCUSSIOX  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Horton.  with  1^  by  1^  by,  ^-in.  angles  with  1  by  x\-in.  lattice,  all  secured  by 
|-in.  rivets,  common  sense  will  indicate  that  the  columns  in  this  group 
should  not  be  used. 

The  "Riile  of  Three"  may  be  accepted  as  an  agent,  to  assist  in  ap- 
proaching the  testing  machine  with  columns  of  a  size  and  cost  so  that 
we  may  hope  for  extended  research.  The  "Rule  of  Three"  may  also 
be  accepted  as  an  agent  to  assist  our  common  sense,  as  shown  in  the 
second  compilation. 

In  the  foregoing,  the  writer  has  attempted  to  point  out  the  de- 
sirability of  using  all  the  rivet  section  possible  in  combining  the  parts 
of  a  composite  compression  member. 

All  research  which  is  available  indicates  that  the  thickness  of  the 
material  in  the  rectangular  compression  member  has  most  to  do  with 
its  efficiency,  thick  material  being  required  for  short  radii  length,  and 
reducing  in  thickness  as  the  radii  length  increases. 

The  piling  together  of  relatively  thin  plates  in  multiple,  with  a  few 
tack  rivets,  and  assuming  that  the  mass  is  homogeneous  is  dangerous. 

Some  comprehensive  proportion  of  stress  through  the  compression 
member  must  be  accepted  as  shear,  and  must  be  provided  for;  if,  on  a 
12  or  15-in.  2-channel  strut  of  medium  size,  practice  dictates  lattice 
of  a  weight  equal  to,  say,  30%  of  the  scantling  weight  of  the  member, 
the  "Rule  of  Three"  will  indicate  that  these  same  relations  must  be  ex- 
tended to  large  or  small  members. 

It  is  not  formulas  that  are  needed  to  extend  our  knowledge  of  the 

compressive  member,  but  comprehensive  research  by  physical  tests. 

Mr.  Shear-        F.  P.  SiiEARWOOD,  M.  Am.  Soc.  C.  E.  (by  letter).— Mr.  Worcester's 

wood,     curve  appears  to  be  more  rational  than  any  of  the  others  he  has  plotted; 

still,    in    common    with    all    column    formulas,    his    assumes    that    the 

flexural  stresses  only  result  from  the  tendency  of  the  member  as  a 

whole  to  bend,  and  no  reduction  is  allowed  for  the  secondary  bending 

from  the  unsupported  component  parts  and  other  unavoidable  bending 

stresses  which  occur  in  many  of  the  compression  sections  now  in  use, 

and  especially   in  those  having   radii   of  gyration   relatively  large   in 

•    comparison  with  their  areas. 

Strict  adherence  to  a  specified  column  formula  has  perhaps  done 
very  much  to  force  designers  to  use  compression  members  which  are 
undesirable  in  nearly  every  way  except  that  they  meet  the  requirements 
of  the  formula  economically  as  regards  material. 

All,  or  nearly  all,  specifications  have  called  for  the  unit  stresses 
in  columns  to  be  determined  solely  by  the  ratio  of  their  length  to  their 
radius,  the  latter  to  be  calculated  from  the  moment  of  inertia  of  the 
section,  without  regard  to  whether  the  lattice  (if  used)  is  capable  of 
developing  it,  or  whether,  in  so  doing,  secondary  stresses  are  induced. 

The  latticed  double  channel  section  with  flanges  turned  out,  so  fre- 
qiiently   used   for   compression    members   of   truss   bridges,    is   a   good 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS  283 

illustration  of  tlie  incompleteness  of  the  ordinary  column  formula,  jyjj.  g^ear 
This  section  is  generally  used  because,  with  a  given  width  from  out  wood, 
to  out  of  chord  gusset  plates,  it  will  give  a  strut  having  the  largest 
radius,  and  therefore  the  highest  permissible  unit  stress;  but,  if  in- 
vestigated, it  will  be  found  that  the  following  stresses  are  almost  in- 
evitable in  such  a  section,  and  of  these  the  column  formula  takes  no 
account,  and  they  are  practically  unprovided  for : 

I. — Stress  due  to  the  flexure  of  the  unsupported  parts  between 
lattice-bar  connections,  which  is  coincident  with  that  due  to  the  flexure 
of  the  column  as  a  whole; 

II. — Stress  due  to  the  eccentricity  of  the  end  connections,  since 
the  center  of  gravity  of  either  channel  is  usually  some  considerable 
distance  from  the  center  of  the  gusset  plates; 

III. — Stress  due  to  the  eccentricity  of  the  lattice-bar  connections; 
for  it  is  usually  impracticable  to  arrange  the  bars  so  that  they  will 
intersect  on  the  center  of  gravity  of  the  channel ; 

IV.- — Serious  but  less  determinate  stresses  are  probably  induced  at 
or  near  panel  points,  where,  owing  to  the  necessary  connections,  it  may 
be  impracticable  to  provide  the  last  few  feet  of  an  important  com- 
pression member  with  either  tie-plates  or  lattice  bars;  and,  even  when 
center  diaphragms  are  provided,  the  continuity  of  the  lattice  system 
is  broken  up,  resulting  in  unknown  bending  moments  in  the  member. 

V. — In  nearly  all  bridges,  the  loads  are  applied  more  or  less  on  the 
inside  of  the  trusses,  thereby  inducing  longitudinal  shear  in  the  sev- 
eral members,  which  in  turn  must  stress  the  eccentrically  connected 
lattice  bars,  and  increase  the  local  stresses,  as  in  III. 

VI.— Lattice  bars,  having  of  necessity  to  be  placed  at  an  aiigle  to 
the  direct  stress  of  the  member,  create  a  distortion  and  bending  when 
the  main  member  is  under  strain. 

Most  of  the  foregoing  defects  are  absent  or  are  minimized  in  mem< 
bers  having  plate  diaphragms,  such  as  H -shapes,  which  are  symmetrical 
about  every  axis.  Secondary  bending  is  largely  eliminated,  and  all 
metal  resists  stress  in  direct  lines  to  the  applied  loads.  They  are  well 
adapted  to  transfer  any  uneven  application  of  load,  but,  unfortunately, 
owing  to  their  relatively  small  radius,  they  cannot  be  made  to  figure  as 
economically  as  flimsy  latticed  members. 

It  seems  to  the  writer  that  columns  with  their  several  parts  tied 
together  with  solid  plates  should  have  more  favorable  consideration 
than  those  which  are  occasionally  tied  together  with  redundant  and 
stress-inducing  lattice  bars. 

Mr.  Worcester,  in  common  with  many  other  authorities,  apparently 
attributes  no  advantage  to  fi:xed  ends  over  pin  ends. 

It  would  seem  reasonable  that  members  with  fixed  ends  should  have 
their  lengths  multiplied  by  some  factor  (say  0.7)  when  the  allowed 
unit  stress  and  limiting  lengths  are  computed. 


284 


DISCUSSION  ON   SAFE   STT?ESSES  IN  STEEL  COLUMNS      [Papers. 


Mr.  Shear- 
wood. 


Fig.  IG  is  given  as  an  example,  and  shows  the  direct  unit  stress 
allowed  by  the  straight-line  formula  of  the  American  Railway  Engi- 
neering and  Maintenance  of  Way  Association,  in  which  only  the 
flexure  of  the  column  as  a  whole  is  considered,  and  also  that  permissible 
if  the  flexure  of  the  unsupported  portions  of  the  individual  leaf  and 
bending  from  the  eccentricity  of  the  lattice  bars  are  also  provided  for. 


Radius  of  {f yi'atiou  =  5.9" 

"        •'        '"         of  single  leaf  =  O.ili" 
Area  of  section  ^  33  sq.  in. 

Allowea  stress  by  straight-line  formula  =  ICOOO-7O7  =1  lT:iO  lli.  per  S'l.  in. 
Permissible  stress  if  some  of  the  other  stresses  due  to  flexure  are  provided  for. 

Unit  stress  ICOOO 

Deduct  for  flexure  of  Column  as  a  whole        4280 
"'        "         "        "    unsupported  leaf  1370 

"        "    bending  stress  from  eccentric 

lattice  connection         [TOO  CS-TO 

9050    Permissible  Stress 

Fig.  16. 

It  seems  probable  that  the  disregarded  secondary  flexural  stresses 
in  latticed  columns  have  caused  the  tests  on  short  lengths  to  disagree 
with  column  formulas  which  are  based  on  the  compression  value  of 
the  metal.  In  devising  a  new  column  formula,  the  many  inherent 
weaknesses  of  latticed  forms  should  be  taken  into  account.  Such  a 
formula  would  have  the  advantage  of  discouraging  the  use  of  ex- 
aggerated forms  with  redundant  metal,  and  encouraging  the  use  of 
members  with  continuously  connected  component  partg. 

Mr.  Rights.  L.  J).  RiGHTS,  Assoc,  M.  Am.  Soc.  C.  E. — This  paper  brings  for- 
ward a  subject  which  is  of  interest,  not  only  to  engineers  who  work 
with  structural  steel,  but  to  all  constructors  who  use  columns  of  other 
material. 

Believing  that  there  is  a  feeling  among  a  number  of  engineers 
that  the  "factor  of  ignorance"  in  regard  to  steel  columns  has  an  in- 
sufficient margin,  the  author  considers  all  the  available  tests,  reduces 
them  to  equivalent  working  values,  and  plots  the  results.  He  then 
breaks  away  fropi  any  attempt  to  assume  a  theoretical  formula,  and 
introduces  a  c^^rve  which  agrees  fairly  well  in  its  middle  portion  with 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS  285 

the  average  of  the  tests,  and  has  the  limiting  values  of  12  000  lb.  at  Mr.  Rights, 
cue  end  and  120      at  the  other. 

As  shown  by  tlie  author's  diagram,  Fig.  1,  most  of  the  formulas 
now  in  common  use  indicate  considerably  higher  values  than  those 
given  by  the  proposed  curve,  and  the  question  at  once  arises  whether 
engineers  are  warranted  in  making  such  a  radical  reduction.  In  the 
light  of  present  knowledge,  the  speaker  does  not  feel  that  it  is  advisable 
to  take  such  a  step.  Many  of  the  tests  which  the  author  has  plotted 
are  from  twenty  to  twenty-five  years  old;  some  of  them  are  on  plain 
shapes,  and  many  of  them,  as  indicated,  are  for  iron.  The  present 
practice  has  been  built  up  from  these  same  tests,  and  is  an  attempt, 
perhaps  in  a  makeshift  way,  to  accommodate  itself  to  the  improved 
conditions  of  manufacture  and  details,  which  have  changed  materially 
since  the  tests  were  made.  It  is  the  speaker's  belief  that  engineers 
would  hardly  feel  justified  in  recommending  this  increased  expense  to 
their  employers  or  clients. 

Although  the  speaker  cannot  agree  with  the  author  as  to  the  large 
reduction  proposed,  nevertheless,  he  feels  that  some  enthusiasts,  over- 
confident in  the  supposed  knowledge  concerning  the  present  state  of 
manufacturing,  have  increased  the  working  values  beyond  safe  limits, 
and  he  would  suggest  that  there  is  a  middle  ground.  He  would  like 
to  offer  as  a  temporary  formula,  and  more  or  less  of  a  compromise,  the 

straight  line  produced  by  15  000  —  75     . 

For  the  initial  point,  15  000  lb.  seems  to  be  a  .satisfactory  value. 
A  very  large  proportion  of  the  steel  now  used  has  an  average  ultimate 
stress  of  60  000  lb.,  with  an  elastic  limit  of  33  000  lb.,  and  the  values 
suggested,  if  properly  reduced,  would  come  within  what  is,  at  present, 
considered  a  safe  limit.  It  will  be  noted  from  the  diagram.  Fig.  17, 
that  this  straight  line  agrees  with  the  author's  plotted  tests  in  the 
middle  portion  fully  as  well  as  the  proposed  curve,  and  that  it  would 
be  practically  tangent  to  the  proposed  curve  at  this  position. 

While  it  is  desirable  to  cut  off  the  column  formula  at  some  higher 

value  of      ,  the  question  also  arises  as  to  where  this  point  shall  be.   Some 
r 

engineers  favor  100,  others,  120,  others,  125,  and  all  of  them  probably 

have  had  occasion  at  some  time  to  consider  columns  at  even  a  higher 

value  of  -  than  125.     When  such  high  values  become  absolutely  neces- 

r 
sary,  the  engineer  uses  the  old  formula  with  discretion.     It  is  the 
speaker's  belief    that,   instead   of  attempting  to  saw  off  the  curve  at 
some  arbitrary  point,  it  should  be  made  fool-proof  by  giving  it  safe 

values  above  200   .  .     It  will  be  seen  that  the   proposed  straight   line 


28G 


DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 


Mr  Rights,  would  end  at  200    ,  and  would  give  fairly  low  values  between  this  and 

r 


120 


I 


In  the  light  of  our  present  knowledge,  or  ignorance,  the  straight- 
line  formula  would  seem  to  be  adequately  accurate  for  all  practical 
use,  and  the  speaker  feels  that  engineers  could  not  do  better  than 
adopt  such  a  simple  formula  until  more  is  learned  about  the  subject. 


\ 

\ 

s 

\ 

-iv. 

^ 

/- 

^-1 

\ 

\ 

N 

N 

J/T''' 

\ 

\ 

s 

s. 

\ 

- 

^ 

V 

X 

\ 

10000 

-^ 

s. 

% 

^% 

\ 

s 

\ 

^ 

^' 

\- 

\ 

8000 

N 

1 

[V 

\ 

\ 

^ 

\ 

^ 

It- 

%, 

\ 

\ 

6000 

\ 

\ 

N 

X 

^ 

^^ 

\ 

X 

^ 

\ 

W\ 

\ 

N 

s 

\ 

■^ 

\ 

N 

N 

s 

2000 

!t\\ 

'■X 

l> 

S 

N 

- 

^^ 

\ 

\ 

0 

\ 

\ 

J 

0        i 

0      ; 

0       1 

u 

jO       i. 

u 

U      i, 

0       < 

0    10 

0     11 

0      1^ 

0        lo 

U      11 

0     10 

0      11 

0    1- 

0 

18 

0     It 

0   200 

DIAGRAM   OF   COLUMN    FORMULAS 
Fig.  17. 

The  speaker's  suggestion  would  be  that  engineers  either  stick  to 
their  present  formulas,  using  them  in  a  conservative  manner,  or  adopt 
some  simple  straight  line,  as  suggested  herein.  New  tests  are  needed, 
not  new  formulas,  and,  until  these  tests  become  available,  it  would  be 
better  for  engineers  to  work  conservatively  along  the  lines  they  have 
been  taught. 

Mr.  Carpenter.  ^     W_    CARPENTER,    AsSOC.    M.    Am.    SoC.    C.    E.     (by    letter).— Mr. 

Worcester  apparently  overlooked  the  fact  that  full-sized  tension  mem- 
bers do  not  develop  the  strength  of  specimen  test  pieces,  and  his  com- 
parison of  the  ultimate  strength  of  full-sized  columns  with  the  ulti- 
mate tensile  strength  of  test  specimens,  therefore,  is  hardly  on  a  fair 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS 


287 


basis.     In  proof  of  the  statement  that  full-sized  tension  members  do  Mr.  Carpenter. 

not  develop  the  strength  shown  by  test  specimens  of  the  same  material, 

the  writer  would  cite  the  tests*  by  J.  E.  Greiner,  M.  Am.  Soc.  C.  E., 

on  built-up  tension  members;  the  "Tension  Tests  of  Steel  Angles  with 

Various  Types  of  End-Connection,"t  by  Frank  P.  McKibben,  M.  Am. 

Soc.   C.  E. ;  and  any  bridge  engineer's  records  of  tests  of  full-sized 

eye-bars. 

Table  11  is  a  summation  of  the  results  of  these  tests  in  form  to 
bring  out  the  point  desired. 

TABLE  11. 


Description  of  tests. 


Specimen  tests. 


J.  E.  Greiner's  tests  of   built-up  ] 
tension  members:   Nos.  3  to  10  of 
Series  A;  Nos.  11  to  18  of  Series 
B:  Average  of  16  tests  of  full- 
sized  members 

J.  E.  Greiner's  tests  of  single-an- 
gle tension  members  connected 
by  both  legs:  Average  of  4  tests. 

McKibben's  tension  tests  of  single 
angles  connected  by  both  legs:    -     „f„„i 
Average  of  12  tests ;  \      sieei. 

Average  of  70  tests  of  full-sized 
eye-bars  reported  in  Mr.  Grei- 
ner's paper    

Average  of  ^  tests  of  eye-bars  10 
by  Its  to  ly%  in.,  made  under  the 
writer's  direction 


g 

OJ  0,0" 


Mild 
steel. 


Mild 
steel. 


Mild 


Medium 

steel.  (?) 


38  aso 


■37580 


3S040 


37  342 


58200 


57280 


56  480- 


63  582 


-  40  187  1  61 630 


Full-sized 

TESTS. 


"Si 
a  &e 


-  o  3 


3  '^si 

s.a'^'.S 
=  iSj-^  P 

j  s  a  3 
-=  oj  s  S 


45  030 


27  450 


Not 
given. 


31270 
30  440 


50  650 

51050 

-  46  750 
57  745 
59160 


Ratio  of 
full-sized 
to  specimen 

TESTS. 


0.73 


0.84 
0.76 


0.87 

0.89 

0.83 
0.91 
0.96 


Note. — All  test  specimens  unannealed,  except  those  of  the  last  item. 

It  should  be  stated  that,  of  Mr.  Greiner's  tests  on  built-up  mem- 
bers, Nos.  1  and  2  of  Series  A  and  all  of  Series  C  were  excluded  from 
■J'able  11  for  the  reason  that  the  members  all  had  defective  (inten- 
tionally so  designed)  end  connections;  likewise,  only  the  angles  which 
were  connected  by  both  legs  in  the  angle  tests  mentioned  were  con- 
sidered; therefore,  the  results  are  the  most  favorable  possible  toward 
the  best  development  of  strength. 

The  high  ratios  of  ultimate  strength  of  eye-bars  to  that  in  speci- 
men tests  must  be  considered  badly  offset  by  the  low  yield-point  ratios. 
The  yield-point  ratios  for  Mr.  Greiner's  angles  are  certainly  not  favor- 

*  Transactions,  Am.  Soc.  C.  E.,  Vol.  XXXVIII,  p.  41. 
t  Engineering  News,  July  5th,  1906. 


288  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Carpenter  able  to  the  tension  side  of  the  ar^iment.  The  other  tests  by  Mr. 
Greiner  were  made  on  members  built  of  small  sections,  which  probably 
accounts  for  the  high  yield-point  values.  Following  the  usual  laws, 
lower  results  for  yield  point  and  (in  lesser  degree)  for  ultimate  strength 
would  be  found  in  the  members  of  truss  bridges  and  other  structures, 
which  are  built  of  thicker  sections. 

It  would  seem  that,  excluding  eye-bars,  it  would  be  unsafe  to  call 
the  average  ultimate  unit  tensile  strength  of  full-sized  tension  mem- 
bers more  than  0.85  times  the  corresponding  strength  of  test  specimens 
of  the  same  material ;  or,  assuming  60  000  lb.  as  the  ultimate  unit 
tensile  strength  of  structural-steel  test  specimens,  the  corresponding 
average  strength  of  full-sized  tension  members  would  be  51 000  lb. 
Now,  put  Mr.  Worcester's  Diagram  of  Column  Formulas  on  the  basis  of 
16  000  .       ^     ,        .        16  000  16  000  „  ,  _  _, 

517)00  ^"^  ''''^  ^"'^    rX^TTO  "^  42^  ^°^-   ''''''^^''   """'^^   ^"'^    '^'' 

value  he  has  chosen  as  the  unit  stress  for  the  ratio,    _  =  0  (12  000  lb.), 

will  become  14  000  lb.  (practically). 

Mr.  Worcester  entirely  disregards  the  effect  of  end  conditions  on 
the  columns  tested,  treating  columns  with  flat  and  hinged  ends  alike. 
This  seems  to  be  rather  unsatisfactory,  since  the  theoretical  influence 
of  the  end  conditions  on  the  strength  of  columns  is  well  backed  up  by 
tests,  and,  undoubtedly,  conditions  arise  which  justify  a  distinction  in 
this  regard..  The  writer  is  wholly  in  favor  of  a  single  formula  *f or 
bridge  work,  and  that  based  on  hinged  ends,  since  this  condition  is 
closely  approximated  in  pin-connected  members,  and  the  strengthening 
effect  of  greater  fixity  of  ends  in  members  with  riveted  connections  is 
offset  in  unknown  degree  by  secondary  stresses  and  unavoidable  eccen- 
tricity of  loading.  It  would  seem  preferable  to  base  a  curve  on,  or 
compare  formulas  with,  full-sized  tests  of  columns  the  end  conditions 
of  which  are  alike.  For  building  work  and  special  cases,  in  which  the 
condition  can  be  unquestionably  realized,  a  formula  for  columns  with 
fixed  ends  would  seem  to  be  entirely  proper. 

Mr.  Worcester  mentions  the  tests  of  Tetmajer,  Marshall,  and 
Christie  as  ''full-sized."  In  his  excellent  paper,  entitled  "The  Prac- 
tical Column  under  Central  or  Eccentric  Loads,"*  Mr.  J.  M.  Moncrieff 
gives  separate  diagrams  covering  all  the  important  series  of  tests  of 
columns  made,  up  to  the  date  of  the  paper,  and  apparently  includes 
all  the  tests  cited  by  the  author.  According  to  these  diagrams,  the 
tests  of  Tetmajer,  Marshall,  and  Christie  were  on  very  small  "full- 
sized"  members,  generally,  such  as  bars  1  in.  square,  small  angles,  and 
other  shapes  and  tubes.  It  seems  that  some  distinction  should  be  made 
between  these  (especially  the  solid  bars  of  insignificant  size)  and  large 

*Transact{ons,  Am.  Soc.  C.  E.,  Vol.  XLV,  p.  334. 


Papers.]      DISCUSSION  ON   SAFE   STRESSES  IN  STEEL  COLUMNS  289 

columns,  such  as  those  tested  by  Bouscaren,  Strobel,  and  the  Water-  Mr.  Carpenter, 
town  Arsenal.     It  was  noted  that  the  particularly  low  result  of  28  000 

lb.   ultimate  strength  for   a  column  having       =  30  was  obtained  in 

the  series  of  "55  tests  at  Watertown  Arsenal  of  3-in.  square  bars  (cold- 
straightened),  mostly  on  pins  IJ  in.  in  diameter,  eight  being  on  pin.s 
from  I  in.  to  2-i  in.  in  diameter,"  all  of  wrought  iron.  It  would  seem 
that  such  tests  should  be  given  very  little  weight  in  this  consideration. 
The  writer  has  failed  to  note  any  test  of  a  properly-constructed,  cen- 
trally-loaded large  column  which  gives  any  such  low  result.  The 
author's  statement  regarding  a  factor  of  only  2  between  the  ultimate 
strength  of  columns  and  a  working  stress  of  16  000  lb.  per  sq.  in.  in 
compression  seems  to  be  misleading,  because,  if  the  16  000  lb.  be  con- 
sidered the  constant  for  reduction  in  one  of  the  usual  column  formulas, 
such  tests  as  have  been  made  on  large  columns  show  an  average  factor 
of  considerably  more  than  2 — perhaps  2.5  for  mild  steel — and  it  has 
been  pointed  out  that  the  average  factor  in  tension  is  about  3.2.  The 
range  of  variation  from  the  average  is  thought  to  be  about  the  same 
in  tension  as  in  compression.  A  careful  study  of  tests  of  large  steel 
columns  leads  the  writer  to  think  that  it  would  not  be  far  wrong  to 
take,  as  the  value  representing  the  ultimate  strength  of  well-propor- 
tioned and  properly-detailed  columns,  in  the  numerator  of  the  Gordon- 
Rankine  or  other  equivalent  column  formula,  41  000  lb.  for  mild  steel 
and  36  000  lb.  for  wrought  iron.  These  values  would  require,  for  equal 
factors  of  safety  based  on  ultimate  strength,  approximately  the  follow- 
ing comparative  values : 

For  tension  in  steel,  16  000  lb.  per  sq.  in. 
For  compression  in  steel,  13  000  lb.  per  sq.  in. 
For  tension  in  wrought  iron,  13  000  lb.  per  sq.  in. 
For  compression  in  wrought  iron,  11  000  lb.  per  sq.  in. 

In  spite  of  the  author's  remarks,  it  seems  difficult  to  get  around 
the  fact  that  engineers  do  and  must  design  with  the  elastic  limit  in 
view,  and  not  the  ultimate  strength,  and  that  the  structure  is  unsafe 
and  possibly  ruined  when  the  elastic  limit  (or  more  properly  perhaps, 
the  yield  point)  is  passed,  in  tension  as  well  as  in  compression.  Also, 
there  is  considerable  strength  beyond  the  yield  point  in  compression, 
which,  as  far  as  it  goes,  is  just  as  valuable  as  the  tensile  strength  be- 
yond that  limit.  There  appears  to  be  a  lack  of  data  on  the  elastic 
strength  of  columns.  Such  as  the  writer  has  been  able  to  find,  indi- 
cate that  the  elastic  strength  will  be  found  below  that  of  test  speci- 
mens, but  not  more  so  than  with  eye-bars  in  tension.  There  also  ap- 
pears to  be  much  greater  danger  from  imperfect  workmanship  and 
injuries  to  material,  in  the  case  of  columns,'  than  in  tension  members, 
for  which  reason  the  writer  agrees  with  the  author,  that  a  lower  unit 


290  DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS      [Papers. 

Mr.  Carpenter,  should  be  iised  ill  compression  than  in  tension,  and  thinks  that  per- 
haps the  ratio  derived  from  the  ultimate  strength  values  proposed,  will 
be  satisfactory,  that  is,  a  compression  unit  of  about  eight-tenths  of  the 
tensile  unit. 

It  is  at  this  point  that  the  writer  would  ask,  speaking  from  the 
viewpoint  of  a  bridge  designer,  why  reduce  the  compression  value? 
Why  not  raise  the  tension  value  ?  Was  not  the  16  000-lb,  unit  chosen 
with  a  view  to  increased  loads,  and  has  not  the  test  of  years  proven 
that  railway  bridges  can  carry,  with  absolute  safety  and  without  ap- 
preciable deterioration,  much  higher  stresses  than  the  equivalent  of 
the  16  000  lb.  ?  If  the  question  of  maximum  loading  is  settled,  the 
writer  sees  only  extravagance  in  designing  a  steel  railroad  bridge  for 
the  tension  unit  stress  of  16  000  lb.,  the  usual  allowance  being  made 
for  impact  and  workmanship  of  the  high  standard  generally  required. 

Neither  experience  with  columns  in  structures,  nor  study  of  tests, 
convince  the  writer  that  there  is  any  cause  for  alarm  in  the  use  of 
the  16  000-lb.  compression  constant  in  working  formulas  for  steel 
columns,  unless  it  be  that  one  cannot  depend  on  having  columns  prop- 
erly proportioned  and  properly  detailed.  An  analysis  of  most  of  the 
large  columns  which  have  been  reported  as  showing  unsatisfactory 
strength  will  show  that  the  columns  were  defective  in  design,  as  com- 
pared with  the  requirements  of  good  modern  practice.  The  writer 
thinks  that  the  principal  trouble,  if  any,  will  be  found  in  the  column 
details,  and  that  if  the  same  attention  is  given  to  the  concentric  ap- 
plication of  loads  and  to  rivet  connections  as  in  tension  members, 
ample  provision  is  made  for  the  full  transmission  of  stress  to  all  parts 
of  members  through  details,  and  the  ratios  of  width  and  length  to 
thickness  are  kept  down  to  the  limits  of  conservative  modern  specifica- 
tions, so  that  columns  will  have  some  body  and  not  be  "built  of  sheet 
iron,"  and  there  will  be  no  failures  nor  cause  for  alarm  with  the 
16  000-lb.  compression  constant  in  columns  of  ordinary  size  and  con- 
struction. 

It  seems  to  the  writer  that,  instead  of  being  cut  off  on  a  horizontal 

line  for  very  low  values  of  ^  ,  say  less  than  20,  a  formula  line  should 
tlieoretically  rise  abruptly  to  the  compressive  limit  of  the  material  at 

=  0.     This,  of  course,  would  make  a  complicated  formula,  and,  as 

r 

such  short  columns  are  unusual,  it  may  be  as  well  to  omit  this  extra 
complication.     It   will   be  noted  that   Mr.   Worcester  omitted  to  plat 

values  for  columns  havincr       <  20,  although  the  serie.s  of  tests  he  men- 

tions  includes  a  large  nunjber  of  such  with  values  which  would  rise  to 
the  limits  of  his  diagram. 

In    conclusion,    the    writer    would    state    that    he    is    opposed    to    a 


Papers.]      DISCUSSION  ON  SAFE  STRESSES  IN  STEEL  COLUMNS  291 

formula  that  is  "chopped  off"  at  the  "long"  end.  Such  a  formula  may  Mr.  carpenter, 
be  all  right  to  design  with,  but  it  is  "no  good"  for  use  in  determining 
the  strength  of  existing  structures.  A  formula  which  best  represents 
the  true  strength  of  a  column,  assuming  its  design  is  correct  and  its 
physical  condition  is  up  to  the  average,  seems  to  be  the  proper  one, 
and  he  does  not  know  of  any  formula  that  fulfills  this  condition  as 
well  as  the  Gordon-Rankine  formula,  in  the  form: 

C 

i  18, 000  \r  / 


Vol    XXXIV.  MARCH,  1908.  No.  3. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED    185  2. 


PAPERS  AND  DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


EFFECT  OF  EARTHQUAKE  SHOCK  ON  HIGH 

BUILDINGS. 

Discnssion.* 


By  Messrs.  Guy  B.  Waite  and  E.  G.  Walker. 


Mr.  Waite.  GrUY  B.  Waite,  M.  Am.  Soc.  C.  E.  (by  letter). — In  the  paper  en- 
titled "Wind  Bracing  for  High  Buildings,"t  the  writer  assumed  a 
horizontal  wind  pressure  of  30  lb.  per  sq.  ft.  as  acting  against  the  en- 
tire windward  side  of  the  structure.  The  stresses  induced  by  the  wind 
force  were  assumed  to  be  resisted  by  the  construction  acting  as  a  canti- 
lever. The  building  was  assumed  to  be  plumb.  No  increased  bending 
moment,  caused  from  an  overhang,  by  wind  pressure  was  thought  neces- 
sary. Provided  the  weight  of  the  building  was  sufficient  to  counteract 
the  overturning  moment  due  to  wind,  the  mass  or  weight  of  the  build- 
ing did  not  enter  into  the  discussion. 

The  resistance  of  the  structure  to  the  horizontal  component  of 
wind  pressure  was  discussed  without  reference  to  whether  it  weighed 
100  lb.  or  1000  000  lb.  Mr.  Chew's  first  conclusion,  that  the  stresses 
produced  by  earthquakes  are  similar  to  those  caused  by  wind,  is  mis- 
leading. While  the  force  from  wind  pressure  is  definite,  and  is  dis- 
tributed throughout  the  structure,  the  force  from  an  earthquake  is  in- 
definite and  unmeasurable,  and  is  distributed  throughout  the  structure 
only  by  acceleration  taken  at  one  place — the  foundation. 

To  compare  the  two  forces,  it  may  not  be  improper  to  liken  them 
to  the  working  of  horizontal  engines  under  a  given  pressure:  the  force 

*This  discussion  (of  the  paper  by  R.  S.  Chew,  Assoc.  M.  Am.  Soc.  C.  E.,  printed  in 
Pnx-ccdiniis  for  January,  1908),  is  printed  in  Procecflinyn  in  order  that  the  views  expressed 
may  lip  l>ioiif:clit  before  all  members  for  further  discussion. 

ri'raii.'^dctions.  Am.  Soc.  C.  E.,  Vol.  XXXIII,  p.  190. 


Papers.]         DISCUSSION  ON  EFFECT  OF  EARTHQUAKE  SHOCK  293 

of  the  wind  being  similar  to  small  pistons  with  long  strokes,  while  the  Mr.  waite. 
force  of  the  earthquake  is  similar  to  a  piston  having  an  indefinitely 
large  area  with  a  very  small  stroke.  In  the  former  case,  the  engine 
force  is  distributed  against  the  side  of  the  building,  is  limited  in 
amount,  and  is  substantially  all  taken  up  by  the  structure ;  in  the  latter 
case,  the  immeasurable  force  from  the  unlimited  piston  area  is  simply 
carried  through  the  foundation  of  the  building,  the  only  force  taken 
up  by  the  building  being  due  to  the  vibration  of  the  foundation. 

On  the  leeward  side  of  a  building  the  force  of  the  wind  is  prac- 
tically nil,  while  the  force  of  the  earthquake  is  substantially  the  same 
on  each  side  of  the  foundation.  The  wind  force  is  all  taken  by  the 
building,  whether  it  be  a  heavy  or  a  light  structure.  The  vibration  of 
the  foundation  of  a  building,  from  earthquake  (other  things  being 
equal),  will  be  the  same,  whether  it  be  heavy  or  light;  the  momentum 
only  will  vary  with  its  weight. 

The  amount  of  the  vibration  of  the  superstructure  of  the  building 
caused  by  earthquake  will  depend  on  the  rapidity  and  length  of  stroke 
of  the  so-called  piston,  the  elasticity  of  the  material  in  the  construc- 
tion, and  the  design  of  the  building.  For  instance,  if  the  foundation 
of  the  building  were  divided  into  two  horizontal  parts,  with  roller  or 
ball  bearings  between  the  parts,  the  earth  vibration  could  pass  through 
v.'ithout  materially  affecting  the  upper  structure,  whether  heavy  or 
light,  the  only  vibration  being  due  to  the  friction  of  the  bearings. 

If  the  vibrations  were  sufficiently  rapid,  and  the  columns  sufiiciently 
long  and  elastic,  probably  little  vibration  would  be  felt,  whether  the 
building  were  light  or  heavy. 

While  some  structural  resistance  is  absolutely  necessary  in  the  case 
of  wind — the  force  being  definite  and  positive — the  structural  resistance 
required  in  the  case  of  earthquakes  will  depend  on  circumstances  and 
design,  the  force  itself  not  being  against  the  building. 

Now,  with  a  properly  designed  building,  having  a  given  mass  and 
a  certain  given  vibration  of  the  foundation,  but  little  vibration  may 
be  caused  to  the  superstructure,  while,  with  more  unfavorable  designs, 
vibration  enough  to  wreck  the  building  might  be  caused,  even  if  the 
best  of  wind  bracing  were  used. 

Buildings  may  be  definitely  braced  against  wind  pressure,  but  they 
cannot  be  definitely  braced  against  earthquakes. 

A  pile  of  brick  may  be  laid  so  that  it  will  resist  wind  pressure,  but 
will  fall  on  account  of  the  acceleration  caused  by  an  earthquake;  this, 
however,  is  no  reason  why  a  sober  fat  man  cannot  stand  up  as  well  as 
a  lean  man  under  both  earthquake  and  wind  pressure.  In  Nature  we 
see  resistances  increasing  with  the  size  and  weight  of  objects;  if  this 
idea  be  carried  out  in  buildings,  we  will  but  obey  the  mathematical 
laws  which  govern  the  stability  of  all  things  composed  of  matter.  The 
heavier  the  building  the  more  horizontal  resistance  it  will  naturally 


294  DISCUSSION  ON  EFFECT  OF  EARTHQUAKE  SHOCK         [Papers. 

Mr.  Waite.  have ;  but  the  writer  does  not  agree  that  the  resistance  should  increase, 
as  indicated  in  the  rational  analysis  from  which  the  author  draws  his 
conclusions. 

It  is  generally  considered  that  additional  weight  helps  to  distribute 
wind  pressures,  and  that  the  force  of  the  wind  is  largely  used  up  in 
frictional  and  other  internal  work  on  the  mass,  to  the  relief  of  specific- 
ally designed  resistances. 

When  a  force  is  set  up  in  a  building,  from  the  vibration  of  the 
mass  composing  the  foundation,  why  should  not  much  of  this  force, 
which  would  otherwise  be  communicated  to  the  braces  and  connections 
of  the  superstructure,  be  lost  in  doing  the  internal  work  referred  to 
above  ? 

As  buildings  can  be  properly  designed  to  resist  wind,  and  can  only 
be  designed  to  escape  the  vibration  of  earthquakes,  it  is  believed  that 
they  should  be  constructed  so  that  their  parts  will  withstand  wind  pres- 
sures, and  that  they  will  then  be  amply  provided  to  withstand  earth- 
quakes. 

Wind  pressures  are  very  frequent,  while  earthquakes  are  very  rare. 
A  building  will  have  need  for  resistance  to  wind  pressures  several 
thousand  times  for  one  possible  resistance  to  earthquake. 

There  is  no  evidence  to  show  that  a  modern  wind-braced  building 
is  not  strong  enough  to  withstand  the  vibration  from  an  earthquake, 
but  there  is  considerable  evidence  to  show  that  it  is  sufficient. 

A  building  well  designed  to  resist  the  force  of  the  wind  should 
have  plenty  of  good-sized  columns.  The  connections  and  braces  of 
these  columns  to  the  girders  should  be  strong  and  positive,  and  there 
•  shoiild  be  the  maximum  depth  of  connections  of  cross-beams  and 
girders  to  columns.  All  constructions  composing  floors,  walls,  parti- 
tions, etc.,  should  be  capable  of  distributing  stresses  and  of  withstand- 
ing vibrations. 

It  seems  to  the  writer  that  reinforced  concrete  fulfills  these  condi- 
tions better  than  any  other  known  material.  In  reinforced  concrete 
the  columns  and  girders  have  a  monolithic  connection  throughout  their 
height,  and,  with  proper  design,  can  take  stress  in  both  directions.  The 
concrete,  being  a  filling  between  the  reinforcing  steel,  can  take  any 
amount  of  vibration  which  will  be  conveyed  by  an  earthquake.  The 
reinforcing  steel  is  run  into  the  columns,  making  a  stronger  connec- 
tion than  possible  in  steel  and  ordinary  fire-proof  construction. 

If  there  be  any  part  of  a  building  in  which  concrete  properly  re- 
inforced cannot  be  designed  as  light  as,  and  perform  the  function  of 
resisting  stresses  better  than,  any  other  fire-proof  material  (in  addi- 
tion to  which  it  preserves  the  steel),  the  writer  is  not  aware  of  it. 
Mr  Walker  E.  G.  Walker,  Jun.  Am.  Soc.  C.  E.  (by  letter). — The  writer  has 
read  with  very  great  interest  Mr.  Chew's  paper  and  the  analysis  with 
which  he  endeavors  to  arrive  at  the  facts  of  the  resistance  of  a  steel- 


Papers.]         DISCUSSION  ON  EFFECT  OF  EARTHQUAKE  SHOCK  295 

framed  building  to  earthquake  shocks.     At  first  sight  it  would  appear  Mr.  Waik-er. 
that  this  subject  is  not  one  which  is  susceptible  of  much  calculation, 
but,  when  the  nature  of  an  earthquake  disturbance  is  considered  more 
closely,  it  at  once  becomes  apparent  that  a  rational  analysis  may  easily 
be  made. 

As  Mr.  Chew  remarks,  the  effect  of  an  earthquake  is  a  wave  motion. 
The  motion  usually  commences  with  small  elastic  earth-vibrations  of 
short  periodicity,  followed  later  by  the  shock  proper,  the  period  of 
which  will  be  much  greater,  from  1  to  2  sec,  after  which  the  vibra- 
tions will  become  slower  and  smaller  until  a  quiescent  state  is  again 
reached. 

•  This  being  the  nature  of  the  force  acting  on  any  structure  during 
an  earthquake  disturbance,  the  writer  does  not  think  that  the  author's 
analysis,  though  correct  and  in  order  as  far  as  it  goes,  is  sufficiently 
extended  to  take  true  cognizance  of  the  initial  conditions  of  the 
problem.  Mr.  Chew  treats  his  building  as  an  elastic  structure,  the 
foundations  of  which  are  subjected  to  an  impact  by  -a  seismological 
wave,  but  he  neglects  the  fact  that  this  wave  is  followed  by  others  at 
intervals  which,  for  a  short  period,  may  be  regarded  as  regular. 

When  an  elastic  body  is  struck  a  blow,  it  tends  to  vibrate  with  its 
own  natural  period  of  vibration,  and  the  amplitude  gets  less  and  less 
until  the  body  comes  to  rest,  the  maximum  distortion,  and  therefore, 
also,  the  maximum  stresses,  occurring  just  after  the  impact.  This  is 
the  state  of  affairs  Mr.  Chew  assumes. 

The  writer  submits,  however,  that  to  treat  the  problem  on  this  basis 
is  insufficient.  The  building  should  be  considered,  not  as  merely  sub- 
jected to  an  impact,  biit  as  acted  upon  by  a  force  the  intensity  of  which 
varies  according  to  a  regular  periodic  law.  The  building  is  then  com- 
pelled to  execute  forced  vibrations,  and,  if  it  should  happen  to  have 
a  natural  period  which  synchronises  with  that  of  the  applied  force,  a 
resultant  vibration  of  very  large  amplitude  would  be  set  up,  causing 
extreme  stresses.  On  the  other  hand,  and  this,  presumably  is  a  more 
common  case,  if  the  natural  period  of  the  building  and  that  of  the 
impressed  force  are  different,  there  would  be  a  continuous  variation  of 
stress  in  all  members  of  the  structure  though  the  range  would  not  be 
so  great. 

The  mathematical  treatment  of  the  problem,  on  these  lines,  though 
a  little  complicated,  only  follows  the  orthodox  method  of  investiga- 
tions into  the  ordinary  problems  relating  to  vibrations.  The  writer  had 
intended  to  present  a  solution  in  some  of  the  simpler  cases,  such  as 
those  dealt  with  in  the  papei',  but,  unfortunately,  the  time  at  his  dis- 
posal has  been  insiifficient  to  take  up  the  question  in  detail.  However, 
it  should  be  possible,  by  treating  the  structure  as  a  vertical  cantilever 
acted  on  by  a  periodic  force,  to  arrive  at  a  law  of  displacement  for  any 
portion,  and  thus  to  get  a  value  for  the  maximum  deflection,  /j ,  at  any 


296  DISCUSSION  ON  EFFECT  OF  EARTHQUAKE  SHOCK         [Papers. 

Mr.  Walker,  point,  as  well  as  an  accurate  knowledge  of  the  range,  or  extreme  values 
of  deflection.  The  step  from  this  to  a  calculation  of  the  stresses  in- 
duced is  easy  and  straightforward. 

With  regard  to  the  author's  first  conclusion,  it  seems  to  the  writer 
that  the  stresses  produced  by  a  shock  will  be  similar  to  those  caused 
by  wind  only  in  cases  where  the  wind  pressure  is  produced  by  gusts 
at  fairly  regular  intervals.  The  increase  or  otherwise  of  stresses  men- 
tioned in  his  second  conclusion  would  be  brought  out  in  an  analysis  on 
the  lines  the  writer  has  mentioned,  and,  in  calculating  the  scantlings 
of  a  new  structure,  the  extreme  fiber  stress  allowed  could  be  settled  in 
accordance  with  the  range  of  stresses  found. 

This  subject  of  the  stresses  induced  in  an  elastic  structure  by 
seismological  disturbances  is  a  very  interesting  and  important  one, 
from  both  practical  and  theoretical  standpoints.  Up  to  the  present, 
rules  and  formulas  have  been  mainly  the  outcome  of  observation  and 
experiment,  rather  than  of  deduction  from  the  theoretical  investiga- 
tion; but  there  is  no  reason  why  the  latter  method  should  not  be  used; 
so  that,  with  a  knowledge  of  the  seismograms  which  have  been  recorded 
from  time  to  time,  it  should  be  possible  to  deduce,  with  a  fair  degree 
of  arccuracy,  the  probable  stresses  which  would  be  induced  by  a  shock, 
and  to  provide  suitably  for  them.  The  writer  only  regrets  that  he  has 
been  unable  to  devote  the  time  necessary  to  work  out  an  analysis  on 
the  lines  he  has  endeavored  to  indicate. 


Vol.  XXXIV,  MARCH,  1908.  No.  3. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEERS. 

INSTITUTED    1852. 


PAPERS  AND   DISCUSSIONS. 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


THE  USE  OF  REINFORCED  CONCRETE  IN 

ENGINEERING  STRUCTURES. 

An  Informal  Discussion.* 


By  Messrs.   M.   S.   Talk,  Eudolph   P.   Miller,   Eugene  W.    Stern, 
AND  H.  C.  Turner. 


M.  S.  Falk,  Assoc.  M.  Am.  Soc.  C.  E. — A  considerable  number  of  Mr.  Faik. 
reinforced  concrete  structures  have  of  late  been  described  with  en- 
thusiasm before  this  Society  and  in  the  technical  press;  and  many,  if 
not  all,  of  the  published  descriptions  make  it  appear  that  these  struc- 
tures have  been  a  complete  success  from  the  time  of  their  inception, 
causing  no  trouble  to  designer,  owner,  or  contractor. 

As  a  rule,  these  descriptions  cover  the  completed  structure  only, 
and  omit  references  to  the  difficulties  and  dangers  encountered  during 
construction. 

Candid  statements  of  facts  in  relation  to  the  use  of  reinforced 
concrete  are  absolutely  necessary  at  the  present  time;  such  statements 
must  be  accurate,  and  should  conceal  nothing,  so  that  they  may  serve 
as  guides  to  others  who  propose  this  construction  for  similar  classes  of 
work. 

Plates  XL  and  XLI  illustrate  the  construction  of  two  buildings, 
entirely  of  concrete,  which  were  built  during  1907,  are  now  in  use, 
and,  to  any  observer,  would  appear  to  be  eminently  satisfactory.  In 
neither  case  will  the  respective  owners  repeat  their  experiences,  since 
in  both  instances  they  have  learned  that  different  methods  would  have 
afforded  structures  which  could  have  been  erected  more  quickly,  at 
less  cost,  and  would  have  been  fully  as  permanent. 

*  Continued  from  February,  1908,  Proceedings. 


298  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  Faik.  Plate  XL  and  Fig.  1,  Plate  XLI,  show  an  ice  storage  house,  the 
outside  dimensions  of  which  are  58  by  92  ft.,  and  with  a  clear  inside 
height  of  42  ft.  8  in.  from  the  top  of  the  basement  floor  to  the  under- 
side of  the  roof  slab.  The  columns  supporting  the  roof  are  18  by  12 
in.  in  cross-section,  and  are  embedded  at  intervals  of  about  11  ft.  in 
the  curtain  walls,  which  are  12  in.  thick  for  the  exterior  and  10  in. 
for  the  interior  walls.  The  building,  which  is  to  store  cakes  of  manu- 
factured ice,  contains  three  chambers  running  the  full  length  of  the 
structure,  the  only  entrance  to  each  being  through  a  small  ice  chute  in 
■  the  front  of  the  building.  There  are  no  windows.  At  first  study,  any 
engineer  would  claim  such  a  structure  to  be  ideal  for  reinforced  con- 
crete; forms  for  vertical  walls  and  for  one  roof  slab  only  were  re- 
quired.   The  history  of  the  case,  however,  refutes  this. 

The  building  was  planned  by  an  architect,  who  called  for  proposals, 
requiring  the  bidding  contractors  to  design  the  reinforced  concrete 
v/ork  subject  to  iiis  approval,  although  he  himself  wrote  the  specifica- 
tions under  which  the  work  was  to  be  built,  and  prepared  preliminary 
plans  showing  his  ideas  as  to  reinforcement.  One  of  the  requirements 
was  that  the  walls  and  columns  were  to  be  designed  to  withstand  an 
assumed  horizontal  thrust  due  to  the  pressure  of  the  ice.  Conse- 
quently, the  columns  were  designed,  by  the  contractors  to  whom  the 
work  was  awarded,  as  vertical  beams  loaded  at  their  centers  with  the 
ice  thrust.  This  explains  (Fig.  1,  Plate  XL)  why  the  reinforcing  rods  in 
the  columns  were  placed  -in  two  lines  parallel  to  the  exterior  faces  of  the 
columns,  instead  of  being  spaced  more  uniformly  throughout  the  cross- 
section.  The  rods  forming  this  reinforcement  ran,  for  the  most  part, 
the  full  height  of  the  biiilding,  and,  ^as  they  were  not  self-supporting, 
it  was  necessary  to  build  a  wooden  structure  to  hold  them  before  any 
concreting  work  was  done.  This  structure  is  shown  in  Fig.  2,  Plate 
XL. 

The  rods  in  the  columns  were  hooped  together  at  short  intervals, 
not  only  by  outside  wires,  but  also  by  wires  crossing  through  the 
center;  moreover,  in  order  to  space  the  corner  column  rods  away  from 
the  forms,  the  contractor  inserted  plate-washers   on  each  corner   rod. 

The  curtain  walls  were  also  reinforced  with  horizontal  and  vertical 
rods  spaced  and  wired  as  shown  in  Fig.  1,  Plate  XLI.  It  is  evident 
that  the  reinforcement  acted  as  a  screen  through  which  the  raw  con- 
crete was  forced  to  pass. 

One  clause  of  the  architect's  specifications  read  as  follows : 

"The  centering  for  columns  shall  not  be  over  half  the  height  of  the 
building  before  concreting  is  commenced  and  for  enclosing  walls  not 
over  10  ft.  in  height,  unless  otherwise  approved." 

Although  the  contractor  should  have  knovm  better,  he  blindly  at- 
tempted to  follow  this  clause.     The  final  results  of  the  work,  taken  in 


PLATE   XL. 

PAPERS,    AM.   SOC.   C.    E. 

MARCH,    1908. 

FALK   ON 

THE    USE  OF 

REINFORCED  CONCRETE. 


FrG.   1.— REINFuHC'lilJ    CO.VCRETE   STRrCTUKE   FOR  ICE  STORAGE. 


Fig.  2  —Scaffolding  to  Support  Reinforcing  Rods. 


Papers.]     DISCUSSION  ON  THE  USE  OE  REINFORCED  CONCRETE  399 

connection  with  the  design,  are  shown  clearly  in  the  photographs,  and  Mr.  Falk. 
require  no  explanation,  except  that  when  the  forms  for  the  lower  por- 
tions of  the  walls  were  stripped,  the  owner,  mistrusting  both  contractor 
and  architect  as  to  the  safety  of  the  work,  called  for  engineering  advice. 

The  structure  was  completed,  after  much  difficulty,  strictly  accord- 
ing to  plan;  dangerous  defects  were  repaired  so  that  no  failure  may 
be  expected,  and  surface  blemishes  were  plastered  so  that  anyone  not 
familiar  with  the  actual  construction  might  believe  the  building  to  be 
an  example  to  be  followed. 

The  building  shown  by  Fig.  2,  Plate  XLI,  was  originally  designed 
by  an  architect  as  a  frame  building  to  be  finished  in  cement  stucco; 
but  a  reinforced  concrete  contractor  convinced  the  owner  that  it  would 
cost  but  little  more  to  make  the  building  entirely  of  concrete,  and  he 
was  given  the  order  to  proceed.  In  fairness  to  the  architect,  it  should 
be  stated  that  he  was  not  consulted  as  to  the  building  after  the 
original  plans  had  left  his  hands. 

When  the  structure  had  reached  about  half  way  to  the  second 
story  the  owner  began  to  suspect  the  character  of  the  work  which  was 
being  done,  and  decided  to  complete  the  building  by  day's  work  in 
charge  of  an  engineer.  No  difficulties  out  of  the  ordinary  were  en- 
countered until  the  I'oof  was  reached. 

The  building  is  40  ft.  square,  and  there  are  four  interior  columns. 
The  roof  is  a  concrete  slab,  sloping  at  about  45°  from  the  horizontal, 
and  is  supported  on  the  side-walls  and  on  two  concrete  beams  running 
the  length  of  the  building  and  carried  by  the  concrete  columns.  The 
concrete  in  the  main  portion  of  the  building  had  been  poured  very  wet; 
but  when  this  mixture  was  placed  on  the  sloping  roof  forms  it  refused 
to  stay  in  place.  Therefore,  wooden  planks  had  to  be  placed  on  top 
of  this  concrete  in  order  to  hold  it  down.  This  method,  however,  was 
exceedingly  difficult,  as  the  roof  was  a  dangerous  place  for  the  work- 
men. The  concrete  was  changed  to  a  drier  mixture,  but  still  re- 
quired the  use  of  the  outside  forms.  As  it  was  impossible  to  lay  very 
much  of  the  roof  in  one  day,  there  were  many  joints.  After  the  con- 
crete had  set  so  that  workmen  could  move  about  without  injury  to  it, 
a  surface  coat  of  mortar,  in  which  was  incorporated  a  so-called  water- 
proof compound,  was  placed.  This  coat  was  colored  with  red  oxide  of 
iron,  so  that  the  final  surface  showed  a  pleasing  red.  The  surface 
coat  was  plastered  smooth,  and  it  seemed  as  though  all  water  would 
be  easily  shed.  The  first  rain  storm,  however,  showed  that  the  roof 
leaked  almost  like  a  sieve.  It  must  be  remembered  that  this  work  had 
been  done  by  day's  labor,  and  not  by  contract,  and  that  there  had  been 
absolutely  no  incentive  for  any  but  the  best  workmanship. 

The  speaker  consulted  several  water-proofing  companies,  asking 
them  to  water-proof  the  roof  without  destroying  the  color  effect  which 
had  been  obtained,  but  not  one  of  these  companies  would  take  the  work 


300  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  Faik.  and  guarantee  it  for  more  than  one  year.  The  use  of  pitch  or  sirailar 
water-proofing  material  was  not  permitted  on  account  of  the  color,  nor 
does  the  speaker  believe  that  any  plastic  material  would  stay  on  this 
roof.  It  was  finally  decided  to  apply  alum  and  soap,  as  in  the  Sylvester 
process,  and  from  its  application  up  to  the  present  time  the  roof  has 
shed  the  rain.  It  has  not  yet  passed  through  both  a  summer  and  a 
winter,  and  it  will  be  interesting  to  note  what  effect  the  temperature 
will  have  on  a  thin  slab  of  this  kind.  The  speaker  would  not  advise 
anyone  to  use  a  reinforced  concrete  roof  of  this  kind. 

Mr.  Miller.  RuDOLPH  P.  MiLLER,  M.  Am.  Soc.  C.  E. — In  the  speaker's  experi- 
ence, along  the  line  of  building  construction,  the  success  of  reinforced 
concrete  in  engineering  work  is  greatly  dependent  on  thorough  and 
intelligent  inspection.  Many  a  good  design  has  been  completely  de- 
feated because  of  the  lack  of  proper  superintendence.  The  materials 
being  used  at  the  present  day  in  this  kind  of  work  are  generally  reliable, 
but  their  improper  handling  has  often  been  responsible  for  poor  re- 
sults. It  is  desired  to  call  attention  here  to  two  defects  that  have  been 
of  too  frequent  occurrence,  which  can  be  avoided  by  a  little  foresight 
in  the  design  and  by  intelligent  supervision  in  the  construction :  First, 
the  displacement  of  the  reinforcement  when  the  concrete  is  placed; 
and  second,  the  formation  of  cavities  in  the  concrete  construction  due 
to  complicated  reinforcement. 

It  would  seem  unnecessary  to  call  the  attention  of  engineers  to  the 
danger  of  the  displacement  of  reinforcing  rods  or  bars  in  reinforced 
concrete  beams.  Concisely  stated,  if  the  displacement  is  upward,  there 
is  a  loss  of  strength  proportionate  to  the  amount  of  displacement;  if 
the  displacement  is  downward,  the  fire-resisting  qualities  of  the  con- 
struction are  impaired,  and  ability  to  resist  fire  is  one  of  the  main 
claims  of  superiority  of  reinforced  concrete  construction.  Judging 
from  experience,  however,  it  seems  to  be  important  to  call  the  atten- 
tion of  engineers  to  the  necessity  of  making  provision  for  preventing 
the  displacement  of  the  reinforcement.  It  is  the  speaker's  opinion 
that,  no  matter  how  carefully  bars  or  rods  are  placed  in  the  moulds, 
or  what  precaution  is  taken  in  the  pouring  of  the  concrete,  there  can 
be  no  assurance  that  the  reinforcement  is  in  its  proper  position  when 
the  work  is  completed,  unless  some  means  have  been  used  to  prevent  a 
movement.  The  only  certain  method  that  has  come  to  the  speaker's 
attention  is  that  used  in  the  so-called  "Unit"  systems,  in  which  all 
the  reinforcing  bars  or  rods  in  a  beam  (and  it  is  equally  applicable  to 
column  construction),  including  the  stirrups,  are  secured  by  heavy 
wire  clamps  or  other  devices  in  such  a  way  that  their  relative  positions 
cannot  alter.  By  using  washers  or  spacers  the  resultant  frame  can  be 
secured  in  the  forms  against  a  bodily  displacement,  and  held  at  a 
proper  distance  from  the  outer  surface  of  the  finished  concrete. 


PLATE   XLI. 

PAPERS,  AM.  SOC.  C.   E. 

MARCH,   1908. 

FALK  ON 

THE   USE  OF 

REINFORCED  CONCRETE. 


Fig.  1.— Reinforcement  of  Walls  and  Columns. 


/ 


ii 


Rf    IR    nr 

■  ■"^'  If 


m        iir      j,r 


|i?   rri 
2  R    I  g  I 


Fig.  i.— Reinforced  Concrete  Building. 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFOECED  CONCRETE 


301 


Besides  assuring  the  correct  position  of  the  reinforcement,  the  use  Mr.  Miller. 
of  the  unit  frames  greatly  simplifies  the  superintendence  of  the  con- 
struction. It  requires  but  a  glance  (comparatively  speaking)  to  see 
whether  all  the  reinforcement  is  in  place  in  the  form  and  whether  the 
proper  frame  is  in  each  form.  The  frame  having  been  built  up  from 
detailed  drawings,  previously  prepared,  the  danger  of  the  omission, 
occasionally,  of  a  bar  or  rod,  of  the  substitution  of  a  bar  of  less  cross- 
section,  or  of  the  use  of  too  short  a  bar,  is  practically  eliminated.  (See 
Fig.  1,  Plate  XLII.) 

The  frames  themselves  may  be  fabricated  at  the  shops  and  shipped 
to  the  job;  or,  if  the  operation  is  sufficiently  large  to  justify  it,  there 
may  be  a  temporary  shop  on  the  premises.  The  particular  advantage  in 
this  is  that  the  forms  can  be  inspected  and  checked  before  they  are  put 
in  place.  A  sample  detailed  drawing  from  which  the  frames  are  made 
is  shown  in  Fig.  1. 


Ul-3J«3J<l-2i4<-i-c->U— i-(;->U2-04>j<-2-0iJ* 3-11— 

f* '■ '■ 8-0-^= ' >[-; 8^1- 

U ■ 21-l-^=— 


0'4^-l-(;-->i-i-l-Git<l'3>|«5^l'3ij 
'^ 8^^^ '■ -] 


No.of 
Pieces 

MATERIAL 

LENOTH 

WEIGHT 

REMARKS 

4 

Rods  1  Vie  Diam. 

24'!" 

4 

.1 

27'4" 

Single  Sockets 

2 

Double  Sockets 

4 

Stirrups  l"x  V6  " 

9'2" 

L    Holes  punched  I  from  each  end 

(i 

8'10" 

L 

4 

8' 9" 

I' 

4 

Ties  l"i  li" 

12" 

t5 

16 

Clamps  with  Bolts 

See  Standard  Sheet  Ho.il. 

Fig.  1. 

Fig.  2,  Plate  XLII,  shows  another  and  quite  satisfactory  method  of 
securing  the  reinforcement  in  position  when  the  style  of  columns  used 
is  such  as  to  admit  of  it.  This  is  a  photograph  of  one  of  the  column- 
girder  connections  in  the  McGraw  Building,  New  York,  recently  de- 
scribed* by  William  H.  Burr,  M.  Am.  Soc,  C.  E. 

The  second  detail  of  construction  which  seems  to  have  escaped  at- 
tention is  the  avoidance  of  too  complicated  a  reinforcement.  In  the 
disposition  of  the  steel,  care  must  be  taken  that  the  several  elements 
are  not  so  closely  spaced  or  so  arranged  as  to  prevent  the  concrete  from 
pouring  between  and  around  them  and  thus  producing  cavities.  The 
size  of  stone  used  in  the  aggregate  should  be  considered  in  connection 
with  the  spacing  of  the  rods,  or  vice  versa.  When  complicated  rein- 
*Proceedmgs,  Am.  t'cc.  C.  E.,  for  October,  19C7. 


302  DISCUSSIOX  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  miler.  forcement  cannot  be  avoided,  the  size  of  the  stones  should  be  reduced 
to  suit  the  condition,  or  the  stone  should  be  eliminated,  and  mortar 
should  be  used.  All  this  applies  particularly  to  column  construction 
and  other  work  where  the  concrete  must  fall  through  considerable 
height.  The  speaker  has  seen  a  column,  the  cross-section  of  which 
was  not  more  than  20%  of  its  embedded  area,  because  of  the  cavities 
formed  by  the  sieve-like  action  of  the  reinforcement.  Ah  instance  of 
what  is  meant,  though  not  as  serious  as  the  case  referred  to,  is  shown 
in  Fig.  3,  Plate  XLII. 

Mr.  Stern.  EuGENE  W.  Stern,  M.  Am.  Soc.  C.  E. — ISTo  Structural  material  in 
recent  years  has  temporarily  won  such  enthusiastic  partisanship,  or 
caught  the  public  eye  to  such  an  extent,  as  reinforced  concrete.  It 
may  be  that  the  reason  for  this  is  that  it  appeals  so  much  to  the 
imagination  of  the  layman. 

It  is  useless  to  consider  all  the  claims  that  have  been  made  for  it; 
but  one  in  particular  should  be  flatly  contradicted,  which  is  that  but 
little  special  knowledge  is  required  in  the  art  of  working  in  this  ma- 
terial, and  therefore  that  it  can  be  largely  done  by  unskilled  labor. 
In  view  of  the  many  fatal  accidents  which  have  resulted  from  the  im- 
proper use  of  this  material,  this  claim  is  not  as  strongly  urged  as  it 
once  was. 

A  matter  of  interest  in  connection  with  the  construction  of  rein- 
forced concrete  work  is  that  contracting  firms,  or  those  who  exploit 
patented  or  deformed  bars,  are  largely  responsible  for  the  designs 
which  go  into  buildings  to-day.  They  submit  their  own  plans,  based 
on  the  use  of  these  bars  or  some  special  method  of  construction,  under 
seme  kind  of  guaranty  as  to  carrying  capacity,  almost  always  without 
any  charge  for  their  services.  It  is  the  speaker's  experience  that  this 
method  leads  to  trouble,  very  often  to  a  lawsuit.  The  client's  interests 
are  supposed  to  be  looked  after  by  the  contractor,  but,  when  any  trouble 
happens,  his  interests,  of  course,  are  forgotten. 

It  is  more  than  ever  necessary,  in  the  use  of  this  material,  that  the 
structural  design  and  supervision  of  the  work  should  be  placed  in  the 
hands  of  competent  professional  engineers  who  represent  the  owner's 
interests  only. 

In  no  other  material  of  construction  is  such  extreme  care  necessary, 
and  such  intelligent,  constant  and  painstaking  supervision  required  in 
every  part  of  the  process. 

Reinforced  concrete  is  a  valuable  material  for  use  in  structures-, 
and  has  a  large  field.  It  has  its  limitations,  however,  and  often,  owing 
to  the  enthusiasm  of  its  advocates,  has  been  used  where  other  ma- 
terials would  have  answered  the  purpose  better. 

As  a  facing  for  buildings  it  has  not  proven  a  success,  for  the  rea- 
son that  it  is  difficidt  to  obtain  a  pleasing  surface,  and  ordinarily  is 


PLATE  XLIl. 

PAPERS,  AM.  SOC.  C.  E. 

MARCH,  1908. 

MILLER  ON 

USE  OF   REINFORCED  CONCRETE. 


Fig.  1.— Unit  System  of  Frame.s. 


Fig  2. — Method  of  Securing  Reinforcement. 


Fig.  3.— Cavities  in  Column. 


Papers.]      DISCUSSION  ON  THE  "USE  OF  REINFORCED  CONCRETE  303 

more  expensive  than  brick.  In  the  construction  of  high  buildings,  the  Mr.  stern, 
speaker  believes  that  it  is  not  as  suitable  for  columns  and  girders  as 
steel-frame  construction.  For  low  buildings,  or  factory  and  mill  build- 
ings, occupying  large  areas  in  outlying  districts,  v/here  there  is  plenty 
of  room  to  handle  the  material,  it  has  proved  a  very  desirable  substitute 
for  mill-constructed  buildings. 

Among  the  favilts  of  reinforced  concrete  work  is  its  tendency  to 
crack,  due  to  the  shrinkage  of  the  concrete.  The  speaker  has  had  to 
deal  with  a  number  of  reinforced  concrete  buildings,  and  none  of  these 
has  been  free  from  cracks  in  various  places.  A  recent  case  was  inter- 
esting: In  a  building  there  were  two  rows  of  colunms  longitudinally, 
dividing  it  into  three  bays.  In  the  center,  between  the  columns,  there 
were  heavy  girders,  but,  in  the  outside  bays,  for  structural  reasons, 
there  were  in  places  light  girders.  The  heavy  girders,  in  shrinking, 
drew  the  columns  slightly  together  and  the  outside  girders  cracked  in 
the  top  flange. 

The  illustrations  shown  by  Messrs.  Talk  and  Miller  are  interesting 
in  showing  what  actually  happens  in  practice.  The  speaker,  however, 
has  seen  much  larger  voids  in  columns  than  any  of  those  illustrated. 
In  one  case,  where  deformed  bars  having  prongs  were  used,  there  were 
voids  in  the  columns  which  practically  occupied  the  whole  area  of  the 
column.  There  was  no  attempt  on  the  part  of  the  contractor  to  scamp 
his  work,  but  the  interlacing  of  the  prongs  formed  a  screen  which  held 
up  the  stone  and  prevented  it  from  becoming  well  consolidated  in  the 
mixture,  with  the  above  result. 

The  use  of  reinforced  concrete  for  railroad  bridges  does  not  seem 
to  be  a  proper  application,  for  the  reason  that  constant  vibration  would 
tend  to  cause  cracks  ultimately,  and  separate  the  reinforcement  from 
the  concrete. 

Where  conditions  would  be  favorable  to  the  rusting  of  steel,  rein- 
forced concrete  is  not  suitable,  unless  cracking  can  be  prevented,  as 
otherwise  the  reinforcement  will  ultimately  rust  out. 

Mathematical  investigations  have  been  carried  to  an  extreme  degree 
of  refinement  in  reinforced  concrete  construction,  and  designs  have 
been  worked  out  on  paper  for  huge  structures  that  stagger  the  imagina- 
tion of  any  but  the  most  enthusiastic.  Only  recently,  a  design  for  a 
bridge  over  Spuyten  Duyvil  Creek,  New  York  City,  has  been  prepared 
by  its  Department  of  Bridges,  involving  the  construction  of  a  rein- 
forced concrete  arch  of  Y03  ft.  span. 

Keinforced  concrete  is  anything  but  an  academic  proposition.  It 
is  eminently  a  practical  one.  Theorists  assume  for  their  computations 
certain  conditions,  some  of  which  may  be  possible,  and  some  of  which 
may  not  be  possible,  to  obtain  in  the  practical  operation  of  construc- 
tion. Many  things  in  the  practical  use  of  this  material  have  yet  to  be 
understood,  and  these  can  only  be  learned  by  experience. 


304  DISCUSSION  ON  THE  USE  OF  EEINFOKCED  CONCRETE     [Papers. 

Mr.  Stern.  Has  the  state  of  the  art,  in  the  use  of  this  comparatively  new  ma- 
terial, progressed  to  such  an  extent  as  to  warrant  the  conclusion  that, 
to-day,  it  is  a  perfectly  safe  and  legitimate  proposition  to  undertake 
to  build  structures,  which  in  magnitude  and  boldness  of  conception  far 
exceed  anything  in  existence  of  similar  type?  Is  it  not  sounder  engi- 
neering to  progress  slowly  along  well-tried-out  lines? 

There  is  wide  difference  of  opinion  as  to  what  the  working  stresses 
ought  to  be,  particularly  in  compression.  The  building  codes  of  the 
various  cities  in  America  are  not  by  any  means  uniform  in  this  re- 
spect, the  allowable  unit  stresses  in  compression  varying  from  350  lb. 
upward,  and  some  engineers  have  recommended  as  high  a  stress  as  750 
lb.  per  sq.  in. 

There  have  been  a  great  many  tests  on  concrete  cubes,  the  data 
obtained  from  which  are  valuable  in  this  discussion.  The  highest  re- 
sults have  been  obtained,  of  course,  where  the  specimens  have  been 
kept  in  moist  sand,  or  submerged  under  water;  but  tests  made  under 
such  ideal  conditions,  which  rarely  obtain  in  practice,  should  not  be 
used  as  a  basis  for  deciding  what  should  be  the  working  stress  of  con- 
crete in  compression,  unless  these  conditions  approximate  those  under 
which  the  structure  itself  is  built. 

Among  the  many  tests  made  at  Watertown  Arsenal,  the  speaker 
would  refer  especially  to  a  series  of  tests  on  12-in.  cubes,  prepared  by 
the  authorities  at  the  Arsenal,  the  results  of  which  are  given  in  their 
Annual  Reports  for  1899  to  1904.  These  blocks  were  allowed  to  set 
in  air,  and  were  stored  in  a  dry,  cool  building  throughout  the  period  of 
the  tests,  which  were  made  after  periods  ranging  from  3  months  to 
5  years.  The  conditions  under  which  the  blocks  were  stored  would  be 
almost  identical  with  those  to  which  reinforced  concrete  work  in  build- 
ing construction  would  be  exposed,  and  these  tests,  therefore,  would 
give  results  more  in  harmony  with  actual  conditions  than  those  in 
which  the  blocks  were  immersed  in  water  or  kept  moist  in  sand  for  a 
number  of  months. 

The  average  of  10  tests,  after  ■>  months,  was  1  958  lb.  per  sq.  in. 

"         "    16      "          "      4        "  "  2  244    " 

"    16      "          •'      1  year,  "  3  330    " 

"    15      "■          "      2  years,  "  2  610    " 

u     -^5       u            u       3        ..  a  2  610    " 

u    -^0      "  "      4      "  "     2  960    - 

a      2      "  "5       "  "     2  630    " 

Disregarding  the  3-month  and  4-month  tests,  the  average  of  58 
tests,  after  1  to  5  years,  was  2  870  lb.  per  sq.  in.  These  blocks  were  all 
made  of  1  part  Alpha  Portland  cement,  2  parts  sand  and  4  parts 
broken  trap  rock,  varying  in  size  in  the  different  specimens  from  i  in. 
to  2^-  in. 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE  305 

It  will  be  noticed  that  the  2-,  3-,  4-,  and  5-year  tests  show  sub-  Mr.  stem, 
stantial  reductions  in  strength  from  the  1-year  tests,  and  the  records 
show,  also,  that  there  was  a  considerable  loss  of  weight,  varying  from 
i  lb.  to  2  lb.  in  each  block. 

There  will  undoubtedly  be  differences  of  opinion  as  to  what  fraction 
of  the  ultimate  strength  should  be  adopted  for  a  safe  working  stress. 
To  compensate  for  the  great  factor  of  ignorance  which  exists  in  the 
construction  of  concrete  and  reinforced  concrete  work,  there  should  be 
an  ample  margin  of  safety.  No  matter  what  care  may  be  taken  with 
the  sampling  and  storing  of  cement,  it  is  practically  impossible,  in  the 
process  of  construction,  to  prevent,  not  only  some  of  the  material 
losing  its  strength,  but  also  to  obviate  defects  in  workmanship. 

It  has  been  considered  for  many  years  that  a  factor  of  safety  of 
from  10  to  20  should  be  used  in  masonry.  The  speaker  sees  no  reason, 
therefore,  why  a  greater  load  than  J  to  ro  of  the  ultimate  strength  of 
laboratory  tests  on  concrete  cubes  should  be  used  in  practice  in  build- 
ing construction,  which  would  give  between  290  and  360  lb.  per  sq.  in. 
as  a  unit  stress. 

Professor  Burr  has  brought  up  the  question  as  to  whether  or  not 
the  Watertown  tests  quoted  by  the  speaker  were  all  made  under  uni- 
form conditions  and  with  the  same  brand  of  cement  and  other  ma- 
terials. As  far  as  can  be  learned  from  the  official  reports  of  these 
tests.  Alpha  Portland  cement  was  used  throughout,  and  the  same  quality 
of  sand  and  stone ;  moreover,  the  specimens  were  stored  under  the  same 
conditions,  throughout  the  years  during  which  the  tests  were  conducted. 

H.  C.  Turner,  Assoc.  M.  Am.  Soc.  C.  E. — During  this  discussion,  Mr.  Turner, 
a  number  of  questions  have  been  raised  which  call  for  an  answer  by 
those  who   are   closely   identified   with  the  construction   of  reinforced 
concrete  buildings. 

In  answer  to  Mr.  Stern's  question  regarding  the  preservation  of  the 
steel  reinforcerbent  in  concrete  structures,  the  following  is  the  experi- 
ence of  the  Turner  Construction  Company  in  razing  a  one-story  build- 
ing, erected  for  the  J.  B.  King  Company,  at  New  Brighton,  Staten 
Island,  in  1902,  which  was  taken  down  during  the  summer  of  1907  to 
make  room  for  a  larger  structure:  The  building  had  reinforced  con- 
crete walls,  9  in.  in  thickness  to  grade  line  and  5  in.  in  thickness  from 
grade  line  to  roof  line,  reinforced  concrete  interior  columns,  11  in. 
square,  and  reinforced  concrete  beams,  girders  and  roof  slab.  The 
foundation  consisted  of  spruce  piling,  cut  off  at  mean  tide  and  capped 
with  reinforced  concrete.  All  steel  reinforcement  was  found  in  per- 
fect preservation  except  a  few  i-in,  hoops  in  the  wall  columns,  which 
were  within  i  to  i  in.  of  the  surface.  These  showed  slight  corrosion, 
which  would  indicate  that  it  is  important  to  secure  all  steel  reinforce- 
ment at  least  |  in.  from  the  exterior  surface.    The  steel  in  the  footings. 


30G  DISCUSSION  ON  THE  USE  OF  REINFORCED  CONCRETE     [Papers. 

Mr.  Turner,  although  alternately  wet  by  the  tide  each  day,  was  in  perfect  condition. 
In  some  cases  this  steel  was  within  |  in.  of  the  surface. 

Numerous  observations  of  a  similar  kind  have  been  made  by  engi- 
neers, and  it  is  now  generally  recognized  that  steel  reinforcement  is 
permanently  preserved  in  concrete  structures. 

It  seems  unfortunate  that  illustrations  of  standard  reinforced  con- 
crete work  have  not  been  shown,  rather  than  those  of  generally  defective 
work,  although  such  illustrations  are  valuable  in  indicating  the  charac- 
ter of  design  and  workmanship  to  be  avoided.  It  must  not  be  assumed, 
however,  that  they  are  typical  of  reinforced  concrete  construction. 
Much  excellent  work  is  being  done  in  New  York  City,  and  in  nil 
cities  in  the  United  States.  Eight  and  ten-story  buildings  are  not 
unusual,  and  it  is  to  be  noted  that  these  buildings  have  proven  espe- 
cially adaptable  for  heavy  storage  or  heavy  manufacturing. 

As  Mr.  Miller  has  stated,  it  is  perhaps  more  difficult  to  secure  good 
workmanship  than  good  engineering  design.  This  is  a  matter  of  or- 
ganization. Good  workmanship  should  be  required,  and  undoubtedly 
can  be  furnished.  There  is  abundant  evidence  of  this,  and  good  work- 
manship costs  but  little  more  than  poor  workmanship.  It  is  necessary 
to  have  a  thorough  and  experienced  organization  of  workmen ;  but  this 
is  just  as  true  in  any  line  of  successful  business. 

Regarding  safe  unit  stresses,  there  is  no  reason  for  a  factor  of 
safety  of  ten.  Reinforced  concrete  buildings  have  a  larger  factor  of 
safety  than  steel  buildings,  because  of  the  monolithic  character  of  the 
construction.  Concentrated  loads  are  distributed  over  larger  areas 
because  the  reinforcement  extends  in  both  directions.  Vibration  is 
largely  reduced.  This  is  well  demonstrated  in  the  Ketterlinus  Build- 
ings, in  Philadelphia.  The  two  buildings  are  about  the  same  size,  8 
stories  in  height;  one  has  a  steel  frame  with  hollow  tile  floors  and 
brick  walls;  the  other,  and  later,  building  has  reinforced  concrete 
columns,  beams,  girders  and  floors,  with  brick  veneer  walls.  Both 
buildings  are  used  for  printing  and  lithographing,  and  are  subjected 
to  practically  the  same  floor  and  machinery  loads.  The  vibration  in 
the  concrete  building  is  very  noticeably  less  than  in  the  steel-frame 
building,  in  fact,  it  can  hardly  be  detected. 

In  the  Robert  Gair  Company  Building,  in  Brooklyn,  there  is  a 
16-ton  embossing  machine  set  on  a  3  by  6-ft.  base  in  the  middle  of  a 
bay  on  the  seventh  floor.  No  deflection  has  occurred  in  the  beams, 
and,  when  the  machine  is  in  operation,  no  vibration  is  perceptible, 
although  the  working  loads  assumed  for  this  building  were  only  200 
lb.  per  sq.  ft. 

Answering  Mr.  Miller's  observatioiis  on  the  value  of  unit  systems 
in  reinforced  concrete  construction,  the  chief  objection  to  them  at 
present  is  the  additional  cost,  which  must  be  paid  by  the  owners. 
Fnit  frames  may   relieve  the  architect  or   engineer  of  some   anxiety 


Papers.]     DISCUSSION  ON  THE  USE  OF  REINFOECED  CONCUKl'l';  307 

and  responsibility,  but  it  is  admitted  that  most  of  the  important  work  M''-  Turner, 
in  the  United  States  has  been  done  with  the  loose-bar  system;  and, 
with  a  proper  organization,  loose  bars,  so-called,  can  be  placed  and 
secured  in  the  work  with  absolute  reliability.  The  owner  looks  for 
results,  and  should  certainly  be  entitled  to  the  difference  in  cost  be- 
tween buildings  constructed  with  loose-bar  systems  and  with  unit 
systems. 


308  MEMOIR  OP   CALVIN   EASTON   BRODHEAD  [Memoirs. 

MEMOIRS  OF  DECEASED  MEMBERS. 

Note.— Memoirs  will  be  reproduced  in  the  volumes  oi:  Transactions.  Any  information 
which  will  amplify  the  records  as  here  printed,  or  correct  any  errors,  should  be  forwarded 
to  the  Secretary  prior  to  the  final  publication. 


CALYIN  EASTON  BRODHEAD,  M.  Am.  Soc.  C.  E.* 


Died  April  29th,  1907. 


Calvin  E.  Brodhead  was  born  in  Pike  County,  Pennsylvania,  on 
December  27th,  1846.  His  family  moved  to  Mauch  Chunk,  Pennsyl- 
vania, in  1851.  He  attended  school  at  vphat  vs^as  knov^n  at  the  time  as 
Park  Seminary,  and  at  St.  Mark's  Parish  School,  where  Felix  Ansart 
was  Principal.  During  vacation  periods,  he  worked  in  the  blacksmith 
shop  of  N.  Eemmel  and  Company,  who  repaired  cars  for  the  old 
Beaver  Meadow  Railroad.  When  the  survey  was  made  for  the  railroad 
from  Bethlehem  to  Bath,  Pennsylvania,  about  1862,  he  found  employ- 
ment on  the  engineer  corps,  and  chose  engineering  as  his  profession. 
After  the  great  flood  of  1862  in  the  Lehigh  River,  which  destroyed  the 
canal  above  Mauch  Chunk,  he  entered  the  service  of  the  Lehigh  Valley 
Railroad,  locating  the  line  over  Wilkes-Barre  Mountain,  between  Penn 
Haven  and  White  Haven.  On  this  work  he  met  the  late  Sidney  Dillon, 
F.  Am.  Soc.  C.  E.,  and  a  friendship  was  formed  which  lasted  until 
Mr.  Dillon's  death. 

After  the  Lehigh  Valley  Railroad  was  opened  to  Wilkes-Barre,  Mr. 
Brodhead  moved  farther  up  the  line  to  what  was  known  as  the  Pennsyl- 
vania-New York  Canal  and  Railroad.  About  1871  he  was  transferred 
from  Wilkes-Barre  to  Bethlehem,  Pennsylvania,  and,  as  Principal  As- 
sistant Engineer,  under  the  late  Robert  H.  Sayre,  Chief  Engineer, 
commenced  locating  the  Easton  and  Amboy  Railroad.  He  remained 
with  the  Lehigh  Valley  Railroad  until  1877.  During  the  building  of 
this  line  (Easton  and  Amboy)  the  construction  of  the  Musconetcong 
Tunnel  was  directly  under  Mr.  Brodhead's  charge. 

From  1877  until  1883  he  was  engaged  in  the  lumber  business,  and 
then  he  formed  a  partnership  with  Lafayette  Lentz,  of  Mauch  Chunk, 
John  Byron  and  Daniel  C.  Hickey,  of  Mt.  Vernon,  New  York,  and 
engaged  in  the  contracting  business.  The  first  contract  of  the  new 
firm  was  for  about  ten  miles  of  very  heavy  work  on  the  Southern 
Pennsylvania  Railway  in  Fulton  County,  Pennsylvania.  In  1885  the 
firm  secured  the  contract  for  the  Vosburg  Tunnel  for  the  Lehigh  Valley 
Railroad.  In  1887  the  firm  of  Brodhead  and  Hickey  succeeded  Lentz 
and  Company,  and  while  connected  with  this  firm  Mr.  Brodhead  was 
engaged  on  several  large  undertakings,  notably  the  Palisade  Tunnel 
for  the  New  York,  Susquehanna  and  Western  Railroad,  and  a  portion 
*Memoir  prepared  by  F.  H.  Clement,  M.  Am.  Soc.  C.  E. 


Memoirs.]  MEMOIR  OF  GEORGE  THOMAS  NELLES  309 

of  the  Pittsburg,  Bessemer  and  Lake  Erie  Railroad,  and  the  Lehigh 
Valley  Railroad.  After  the  death  of  Mr.  Hickey,  in  1894,  the  firm 
name  was  changed  to  C.  E.  Brodhead  and  Brother,  and  subsequently 
to  the  Brodhead  Construction  Company,  under  which  name  the  firm 
continued  work  until  Mr.  Brodhead's  death.  It  was  in  the  contract 
business  that  Mr.  Brodhead  spent  the  most  active  part  of  his  life,  and 
in  that  he  was  best  known  and  most  successful.  He  was  a  man  of 
quick  ideas,  and  was  a  born  locating  engineer,  in  which  capacity  he 
was  frequently  called  in  consultation. 

Mr.  Brodhead  continued  to  be  interested  for  many  years  in  the 
coal  and  lumber  bvisiness,  having  large  interests  in  Kentucky.  He 
was  twice  married,  and  three  children  by  his  first  wife  survive  him. 

Mr.  Brodhead  was  elected  a  Member  of  the  American  Society  of 
Civil  Engineers  on  February  21st,  1872. 


GEORGE  THOMAS  NELLES,  M.  Am.  Soc.  C.  E. 


Died  November  15Tir,  1907. 


George  Thomas  Nelles,  son  of  George  W.  Nelles  and  Virginia  Hobbs 
Nelles,  was  born  on  April  15th,  1856,  in  Muscatine,  Iowa. 

His  boyhood  was  spent  in  Leavenworth,  Kansas,  to  which  place  his 
parents  moved  in  the  summer  of  1857.  Mr.  Nelles  prepared  for  college 
in  the  private  school  of  the  Reverend  (now  Bishop)  John  Mills  Ken- 
drick,  and  was  graduated  from  the  Rensselaer  Polytechnic  Institute 
with  the  degree  of  C.  E.  in  June,  1877. 

After  a  few  months'  work  as  instrumentman  with  the  United  States 
Engineer  Corps  at  Leavenworth,  he  entered  the  service  of  the  Kansas 
City,  St.  Joseph  and  Council  Bluffs  Railway,  as  Assistant  Engineer 
in  charge  of  surveys  and  relocation. 

In  the  summer  of  1878  he  re-entered  the  Government  service,  as 
United  States  Assistant  Engineer,  on  the  Missouri  River  improvement, 
having  in  charge  at  various  periods  the  work  at  Atchison,  St.  Joseph, 
and  Leavenworth,  until  the  spring  of  1883  when  he  was  elected  City 
Engineer  of  Leavenworth,  Kansas.  Entering  upon  his  duties  at  a  time 
when  the  city  was  growing  rapidly,  he  directed  much  public  work, 
supervising,  during  his  term  of  office,  the  expenditure  of  more  than 
$1  250  000  in  grading  and  paving  streets  and  constructing  sewers,  cul- 
verts, and  bridges. 

During  his  term  of  six  years  as  City  Engineer,  Mr.  Nelles  was  also 
Consulting  Engineer  for  the  Western  Home  for  Disabled  Volunteer 
Soldiers;  Chief  Engineer  of  the  Riverside  Coal  Company;  Chief  En- 

♦Memoir  prepared  by  F.  E.  Bissell,  M.  Am.  Soc.  C.  E. 


310  MEMOIR  OF  GEORGE  THOMAS  NELLES  [Memoirs. 

gineer  of  the  Leavenworth  Rapid  Transit  Company;  and  Chief  Engi- 
neer of  the  East  Omaha  Land  Company, 

At  the  organization  of  the  Nebraska  and  Colorado  Stone  Company, 
in  1889,  Mr.  Nelles  became  its  Secretary  and  Manager.  The  company 
operated  quarries  in  Nebraska  and  Colorado,  contracting  not  only  to 
furnish  stone,  but,  also,  in  many  cases,  for  the  complete  erection  of  the 
structure. 

Severing  his  connection  with  the  Stone  Company  in  1891,  he  en- 
tered the  general  contracting  business,  constructing  sewers,  pavements, 
bridges,  water-works  and  river  and  harbor  improvements.  The  largest 
and  most  important  contracts  handled  during  the  four  years  he  spent 
in  this  work  were  the  construction  of  the  sewers  in  Denver,  Colorado, 
and  the  harbor  improvements  in  the  Mississippi  River  at  St.  Louis, 
Missouri. 

In  the  spring  of  1895  Mr.  Nelles  again  entered  the  Government 
service  as  U.  S.  Assistant  Engineer,  on  the  Tennessee  River  improve- 
ment at  Chattanooga,  Tennessee.  During  his  six  years  of  service  on 
the  Tennessee  River  and  its  tributaries,  many  important  and  difficult 
problems  presented  themselves.  He  made  a  careful  study  of  the  con- 
struction of  locks  and  dams  under  the  conditions  of  fluctuating  velocity 
of  current  and  volume  of  discharge  which  there  prevail.  His  reports 
on  all  subjects  assigned  to  him  were  always  exceedingly  full  and  com- 
plete. He  prepared  detailed  tables  showing  the  cost  of  construction  of 
the  lift  and  gtiard  locks  at  Colbert  Shoals,  Alabama.  He  investigated 
the  discharge  of  the  Tennessee  River,  checking  the  formulas  with  the 
actual  measured  velocities,  and  determining  for  this  stream  the 
value  of  n,  in  Kutter's  formula.  His  solution  of  the  problem  of  the 
effect  of  a  dam  on  a  submerged  discharge,  and  on  the  surface  level 
of  the  upper  pool,  reached  in  his  study  of  projects  for  the  improve- 
ment of  that  part  of  the  Tennessee  River  known  as  the  "Suck,"  is  a 
material  addition  to  engineering  knowledge. 

Mr.  Nelles  studied  the  conditions  on  the  French  Broad  River,  and 
made  plans  for  widening  and  deepening  the  channel ;  he  examined  and 
reported  on  the  necessity  of  making  any  improvement  of  Powells 
River;  made  plans  and  estimates  for  the  low-water  improvement  of 
the  Hiawassee,  Little  Tennessee,  and  Clinch  Rivers,  and  also  reported 
on  the  feasibility  of  making  improvements  on  the  Holston  River. 

The  same  careful  attention  to  details,  and  a  comprehensive  con- 
sideration of  all  the  component  parts  of  the  subject,  characterize  each 
of  these  reports.  They  show  that  rare  combination,  complete  theoreti- 
cal knowledge  and  practical  ability. 

In  June,  1901,  Mr.  Nelles  was  transferred  to  Cleveland,  Ohio,  as 
U.  S.  Assistant  Engineer  in  charge  of  the  improvements  of  the  harbors 
on  Lake  Erie  at  Cleveland,  Lorain,  and  Fairport.  The  same  thorough- 
ness and  attention  to  detail,  combined  with  indomitable  energy  and 
great  administrative  ability,  characterized  his  work  there. 


Monioirs.]  MEMOIR   OF   HERBERT   FRANKLIN   NORTHRUP  311 

The  earnestness  with  which  he  worked,  the  ability  which  he  brought 
to  the  work,  and  the  honesty  of  his  dealings,  combined  with  his  cheer- 
ful disposition,  made  him  a  very  companionable  man,  both  socially  and 
professionally. 

His  health  began  to  fail  in  1903.  Two  surgical  operations  failed  to 
give  more  than  temporary  relief,  and  he  died  at  Cleveland,  Ohio,  on 
November  15th,  1907. 

On  February  15th,  1884,  Mr.  Nelles  was  married  to  Miss  Lena 
Ralston,  who,  with  one  son,  survives  him. 

Mr.  Nelles  was  a  Member  and  a  Director  of  the  Civil  Engineers' 
Club  of  Cleveland.  He  was  elected  a  Member  of  the  American  Society 
of  Civil  Engineers  on  October  3d,  1888,  and  contributed  to  the  Trans- 
actions  a  discussion*  on  the  paper  by  the  late  George  W.  Rafter,  M. 
Am.  Soc.  C.  E.,  entitled  "On  the  Flow  of  Water  over  Dams ;"  and  also 
a  discussionf  on  the  paper  by  the  late  R.  C.  McCalla,  M.  Am.  Soc. 
C.  E.,  entitled  "Improvement  of  the  Black  Warrior,  Warrior,  and 
Tombigbee  Rivers,  in  Alabama." 


HERBERT  FRANKLIN  NORTHRUP,  M.  Am.  Soc.  C.  E.  | 

Died  January  21st,  1908. 


Herbert  Franklin  Northrup,  born  on  a  farm  near  Shoreham,  Ver- 
mont, on  October  9th,  1850,  was  the  youngest  child  of  Nazro  and  Mary 
Hawes  Northrup. 

After  attending  the  village  school  he  prepared  for  college  in  Kim- 
ball Union  Academy,  and  entered  Middlebury  College,  Vermont,  in 
the  class  of  '73.  He  next  taught  mathematics  and  English  for  two 
years  at  a  boys'  .school  in  Flushing,  Long  Island.  He  then  took  a 
graduate  course  in  engineering,  in  Sheffield  Scientific  School,  Yale,  in 
the  class  of  '78. 

His  first  engineering  engagement  was  upon  the  Lake  Champlain 
breakwater  at  Swanton,  Vermont,  in  1878,  and  in  1879  he  was  engaged 
on  railroad  maintenance  work  at  Salem,  Massachusetts.  In  the  spring 
of  1880  he  entered  the  employ  of  the  Texas  Pacific  Railroad  as  As- 
sistant Engineer  of  construction,  and  was  located  at  Fort  Worth,  Texas. 
He  continued  in  .the  employ  of  that  company,  in  responsible  positions, 
until  the  completion  of  its  construction  in  1885. 

On  February  2d,  1882,  he  was  married  to  Miss  Cornjelia  F.  Allan, 
of  New  Haven,  Connecticut. 

*  Transactions,  Am.  Soc.  C.  E.,  Vol.  XLI V,  p.  359. 
t  Transactions.  Am.  Boc.  C  E.,  Vol.  XLIX.  p.  384. 
t  Memoir  prepared  by  J.  J.  McVean,  M.  Am.  Soc.  C.  E. 


312  MEMOIR  OF  HERBERT  FRANKLIN   NORTHRUP  [Memoirs. 

He  entered  the  employ  of  the  Missouri  Pacific  Railroad  Company 
in  1885,  as  Assistant  Engineer  of  construction  in  Missouri  and  Kansas, 
and  in  September,  1886,  he  engaged  with  W.  V.  McCracken  and  Com- 
pany, Contractors,  as  Chief  Engineer  on  the  construction  of  railroads 
in  Ohio  and  Indiana.  From  August  to  November,  1887,  he  was  en- 
gaged as  Locating  Engineer  on  the  Duluth,  South  Shore  and  Atlantic 
Railway,  in  the  northern  peninsula  of  Michigan. 

In  November,  1887,  the  writer  engaged  him  as  engineer  in  charge 
of  preliminary  and  location  surveys  for  the  Chicago  and  West  Michi- 
gan Railway,  from  Baldwin  to  Traverse  City,  Michigan,  75  miles, 
which  was  finished  in  June,  1888.  He  was  then  engaged  until  June, 
1889,  upon  some  construction  work*  in  the  East,  when  he  again  returned 
to  take  charge  of  the  construction  of  the  road  from  Baldwin  to 
Traverse  City,  following  which  he  had  charge  of  the  location  and  con- 
struction of  an  extension  of  about  90  miles  from  Traverse  City  to 
Petoskey,  Michigan,  which  was  completed  in  1893.  In  1893  and  1894 
he  was  engaged  with  the  Detroit,  Bay  City  and  Alpena  Railroad,  and 
from  1895  to  1901  was  in  private  practice  and  City  Engineer  of 
Traverse  City,  Michigan,  and  designed  a  proposed  water  supply  for 
that  city.  During  this  time  he  also  located  several  miles  of  road  for 
the  Lake  Superior  and  Ishpeming  Railroad  Company. 

In  1902  he  entered  the  employ  of  the  Cleveland,  Cincinnati, 
Chicago  and  St.  Louis  Railroad  Company,  in  charge  of  a  residency  on 
the  relocation  and  construction  of  its  line  for  double  track,  and  reduc- 
tion of  grades  and  curvature,  where  he  had  charge  of  some  very  diffi- 
cult and  heavy  work,  especially  the  construction  of  several  large-span 
concrete  arches. 

In  May,  1905,  he  formed  a  partnership  with  the  writer,  as  Consult- 
ing Engineers,  with  office  at  Grand  Rapids,  Michigan,  where  he  was 
engaged  until  his  death. 

Mr.  Northrup  was  beloved  by  all  who  knew  him.  He  was  of  a 
very  modest  and  retiring  disposition,  amiable,  a  staunch  friend,  and  a 
thoroughly  honorable  business  man.  Quiet  and  even-tempered,  honest 
in  all  his  dealings,  he  had  not  only  the  entire  confidence  of  his  em- 
ployers, but  also  the  love  and  friendship  of  his  assistants. 

One  of  his  many  assistants,  now  occupying  a  responsible  position 
with  the  City  of  Buffalo,  says  "I  knew  him  as  a  gentleman  and  an  en- 
gineer, and  nothing  can  be  added  to  that.  His  even  temper  and  kindly 
ways  always  left  a  pleasant  recollection." 

His  death  was  very  sudden  and  unexpected;  after  a  severe  fall  on 
an  icy  sidewalk,  he  was  attacked  with  prostatitis,  necessitating  an 
operation  from  which  he  did  not  rally. 

Mr.  Northrup  was  a  Royal  Arch  Mason,  a  member  of  the  Delta 
Kappa  Epsilon  College  Fraternity,  and  was  elected  a  Member  of  the 
American  Society  of  Civil  Engineers  on  January  6th,  1892. 


Memoirs.]  MEMOIR  OF  WILLIAM  ROBERTS  313 

WILLIAM  ROBERTS,  Assoc.  Am.  Soc.  C.  E.* 


Died  December  28th,  1907. 


William  Roberts  was  born  in  Watertown,  Massachusetts,  on  March 
25th,  1835.  He  was  a  son  of  John  Roberts,  a  descendaoit  of  one  of  the 
old  Boston  families.  His  x^arents  moved  to  Waltham  soon  after  his 
birth,  and  he  attended  the  Waltham  Public  Schools,  the  private  school 
of  Daniel  French,  and  the  Allen  School  of  Newton. 

He  entered  the  employ  of  the  Boston  Manufacturing  Company,  in 
the  machine  shop,  as  a  start  toward  the  development  of  his  mechani- 
cal genius.  Obtaining  permission  from  the  Fitchburg  Railroad,  he 
ran  an  engine  from  Waltham  to  Boston  a  number  of  times.  He  then 
went  to  Virginia,  where  he  studied  at  the  establishment  of  the  Norfolk 
Manufacturing  Company.  When  very  young  he  entered  the  United 
States  Navy.  He  was  Third  Assistant  Engineer  under  Commodore 
Perry  when  he  opened  the  Ports  of  Simoda  and  Hakodadi,  in  Japan, 
and  served  on  the  Michigan,  on  the  Great  Lakes  in  1856,  and  on  the 
steam  frigate  Roanoke,  on  the  Coast  of  Central  America  in  1857.  He 
was  one  of  the  officers  on  the  steamer  Fulton  which  captured  Walker, 
the  filibuster,  in  1858,  and  he  served  on  the  Memphis  on  the  Paraguay 
expedition  in  1859. 

In  July,  1858,  Mr.  Roberts  was  promoted,  becoming  Second  As- 
sistant Engineer,  and  one  year  later  he  was  made  First  Assistant.  He 
resigned  in  September,  1859,  but,  in  response  to  his  country's  call,  re- 
enlisted  in  the  Navy  in  April,  1861.  In  1863  he  became  Chief  Engi- 
neer. 

During  the  attacks  on  the  forts  and  batteries  at  Pensacola  Bay,  in 
1861,  he  was  on  the  frigate  Niagara;  the  steam  sloop  Housatonic 
carried  him  to  a  point  off  Charleston,  in  1862,  when  she  drove 
two  iron-clad  rams  into  port.  He  was  attached  to  the  frigate  Niagara 
repairing  at  Charlestown  Navy  Yard,  during  1863  and  1864. 

After  his  retirement  from  the  Navy  he  returned  to  Roberts'  Cross- 
ing, Waltham,  and  joined  his  father  in  the  manufacture  of  paper,  and, 
even  after  his  father's  death,  he  carried  on  the  business  under  the  firm 
name,  John  Roberts  and  Son. 

The  manufacture  of  roofing  paper  was  the  principal  product  of  the 
mill  until  his  ever-active  mind  turned  to  the  then  new  article,  asbestos, 
and  his  mill  was  the  first  to  produce  asbestos  fire-proof  paper,  the  secret 
of  the  process  being  held  by  him  for  many  years. 

He  declined  the  acceptance  of  public  office,  notwithstanding  the 
many  entreaties  on  the  part  of  his  friends.  The  only  State  positions 
he  held  were   Commissioner  on  Prisons,   and   Representative  to   the 


*Menioir  prepared  by  Sumner  Milton,  Esq. 


314  MEMOIR  OF  WILLIAM  ROBERTS  [Memoirs. 

General  Court.  He  was  a  staunch  Republican,  and  was  sent  as  a 
delegate  to  the  State  Convention  for  many  years.  He  was  a  Member 
of  the  Waltham  Board  of  Cemetery  Commissioners,  and  a  Director  of 
the  Waltham  National  Bank. 

Mr.  Roberts  was  a  life  member  of  Monitor  Lodge,  A.  F.  and  A.  M., 
and  Post  29,  G.  A.  R.,  of  Waltham.  He  belonged  to  the  Military  Or- 
der of  the  Loyal  Legion  of  the  United  States,  also  the  American  So- 
ciety of  Mechanical  Engineers. 

"He  serves  God  well,  who  serves  his  creatures,"  truly  speaks  the 
life  of  William  Roberts.  Never  was  he  known  to  refuse  to  help  a 
worthy  person  or  project.  Many  leave  public  bequests  and  are  thought 
generous,  but  Mr.  Roberts'  method  was  to  give  continuously;  and,  as 
was  his  nature,  quiet,  just,  liberal,  honest,  and  philanthropic,  so  was 
his  giving,  and  there  are  many  individuals  and  institutions  who  miss 
his  beneficence. 

Mr.  Roberts  was  an  interesting  conversationalist,  having  toured  the 
world.  He  was  especially  interested  in  the  ocean,  and  crossed  the  At- 
lantic on  all  the  finest  new  steamers,  his  knowledge  of  mechanical  en- 
gineering enabling  him  to  note  all  the  latest  improvements  in  the 
engines.  It  was  difficult,  indeed,  to  ask  a  question  on  country  or  prod- 
uct on  which  Mr.  Roberts  could  not  give  valuable  information,  and  in 
such  a  simple  manner  that  a  child  could  enjoy  his  talk. 

On  October  27th,  1879,  Mr.  Roberts  married  Eva  C,  daughter  of 
Hon.  Gideon  Haynes,  and  their  home  was  always  at  Waltham,  His 
married  life  was  one  of  devotion,  and  it  would  be  difficult  to  decide 
whether  the  palm  should  be  given  to  him  or  to  his  companion  in  life. 

William  Roberts  was  elected  an  Associate  of  the  American  Society 
of  Civil  Engineers  on  June  4th,  1884. 


If  1-?  % 


>»>;», i:>;i.;li»il,»i'i»;. 


;  M' 


If   14   i«  »« 

i>  n  !#. 

i^  t*  u  1»  t 

1.    ft'^ir'-  Ik'   aw  '  A^ 


^TSSING 


VOLUME  S  ) 

ISSUE#  

MONTH 

YEAR  I'^Oh 


r 


%  m 


^-.i  •M»'j)!;ji5J}»J»«.M}«M".M>i'1j 


[«  It  Hw 

iv  i»  «^  1 


Vol.    XXXIV 


MAY,  1908. 


No.  8. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEEES 

1  N  S  T  1  T  L'  T  E  L)    18  5  3 


PAPERS  AND   DISCUSSIONS 

riiis  Sofiety  is  n'>t  i-espousible,  as  a  body,  for  the  facts  aud  opinions  advaiict»d 
in  any  of  its  publications. 


CONTENTS 

Papers :  paoe 

I'lirve  Resistance  in  Water  Pipes. 

By  Ernest  W.  Schoder,  Assoc.  M.Am.  See.  C.  E 416 

Not«s  Upon  Docks  and  Harbors. 

By  Li'THER  Wagoner,  M.  Am.  Soc.  C.  E 446 

Discussions  : 

The  Flood  of  March,  1907.  in  the  Sacramento  and   San  Joaquin   River  Basins, 
California. 
By  Messrs.  Luther  Wagoner,  H.  H.  Wadsworth  and  George  L.  Dillman 460 

Erection  of  the  Bellows  Falls  Arch  Bridge. 

By  F.  W.  Skinner,  M.  Am.  Soc.  O.  E 46H 

Safe  Stresses  in  Steel  Columns. 

By  William  Cain,  M.  Am.  Soc.  C.  E 477 

The  ElectriflcatioD  of  the  Suburban  Zone  of  the  New  York  Central  and  Hudson 
River  Railroad  in  the  Vicinity  of  New  York  City. 
By  Messrs.  Edwin  B.  Katte.  W.  S.  Murray.  George  A.  Harwood,  W.  B.  Pot- 
ter, Frank  J.  Sprague,  Henry  G.  Stott,  and  William  J.  Wilgus 484 

Recent  Developments  in  Pneumatic  Foundations  for  Buildings. 

By  Messrs.  F.  W.  Skinner,  T.  Kennard  Thomson,  and  Louis  L.  Brown 521 

Memoirs: 

Charles  Haynes  Haswkll,  Hon.  M.  Am.  Soc.  0.  E 534 

James  Dun.  M.  Am.  Soc.  C.  E 537 


XLVI. 

XLVII. 

XLVIIl. 

XLIX. 

L. 

LI. 

LIl. 

LIU. 

LIV. 

LV. 

LVI. 

LVII. 

LVIII. 

LIX. 

LX. 

LXI. 

LXII. 

LXIII. 

LXIV. 

LXV. 


PLATES 

Experimental  Pipe  Line  at  Cornell  University  Hydraulic  Laboratory;  Nozzle 

and  Curves 419 

Diagram  Showing  Results  of  Experiments  on  Curve  Resistance  in  Water 

Pipes 423 

Statistical  Chart  Showing  the  Progress  of  Banking,  Commei-ce,  Shippintf, 

Population,  etc 449 

Erection  of  Eads  and  Washington  Bridges 469 

Erection  of  Niagara  Railway  and  Highway  Arches 471 

Erection  of  the  Kaiser  Wilhelm  Bridge 473 

Erection  of  the  Garablt  Viaduct  and  the  Bonn  and  Diisseldorf  Bridges 475 

Erection  of  the  Fairmount  Park  and  the  Rochester  Driving  Park  Bridges. .  477 

Erection  of  the  Lake  Street  and  Panther  Hollow  Bridges 477 

Erection  of  the  Kornhaus  Bridge 477 

Plans,  etc.,  of  Main  Steam  Piping  in  Power-Stations 485 

General  Diagram  of  Connections,  Poit  Morris  Power-Station 487 

Plans  of  Port  Morris  Swit<h-House 489 

Sections.  Port  Morris  Switch- House 491 

Wiring  Diagram,  Typical  Sub-Station 493 

Diagram  of  Positive  Feeders,  N.  Y.  C.  &  H.  R.  R.  R.  Electrification 495 

Cross-Section.  Removal  of  Grand  Central  Train-Shed 497 

Traveler  for  Removal  of  Grand  Central  Train  Shed ' 499 

Views  Showing  Removal  of  Grand  Central  Train-Shed .  501 

Appliances  used  in  Caissons  for  Foundations  of  Buildings 523 


V^ol.    XXKIV  MAY,  1908.  No.  S. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEERS 

1  N  S  T  1  T  I'  T  K  I)    1  8  .5  -2 


PAPERS  AND   DISCUSSIONS 

This  Society  is  ii'it  respansible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


CONTENTS 

Papers :  page 

Curve  Resistance  in  Water  Pipes. 

By  Ernest  W.  Schoder,  Assoc.  M.  Am.  Soc.  C.  F, 416 

Notes  Upon  Docks  and  Harbors. 

By  Luther  Wagoner,  M.  Am.  Soc.  C.  E 446 

Discussions : 

The  Flood  of  March,  1907.  in  the  Sacramento  and   San  Joaquin   River  Basins, 
California. 
By  Messrs.  Luther  Wagoner,  H.  H.  Wadsworth  and  George  L.  Dillman 460 

Erection  of  the  Bellows  Falls  Arch  Bridge. 

By  F.  W.  Skinner,  M.  Am.  Soc.  C.  K 468 

Safe  Stresses  in  Steel  Columns. 

By  William  Cain,  M.  Am.  Soc.  C.  E 477 

The  ElectriflcatioD  of  the  Suburban  Zone  of  the  New  York  Central  and  Hudson 
River  Railroad  in  the  Vicinity  of  New  York  City. 
By  Messrs.  Edwin  B.  Katte.  W.  S.  Murray,  George  A.  Harwood,  W.  B.  Pot- 
ter, Frank  J.  Sprague,  Henry  G.  Stott,  and  William  J.  Wilgus 484 

Recent  Developments  in  Pneumatic  Foundations  for  Buildings. 

By  Messrs.  F.  W.  Skinner,  T.  Kennard  Thomson,  and  Louis  L.  Brown 531 

Memoirs : 

Charles  Haynes  Haswell,  Hon.  M.  Am.  Soc.  C.  E 534 

James  Dun.  M.  Am.  Soc.  C.  E 537 

PLATES 

XL VI.  Experimental  Pipe  Line  at  Cornell  University  Hydraulic  Laboratory;  Fozzle 

and  Curves 419 

XLVII.  Diagram  Showing  Results  of  Experiments  on  Curve  Resistance  in  Water 

Pipes 423 

XLVIIl.  Statistical  Chart  Showing  the  Progress  of  Banking,  Commerce,  Shipping, 

Population,  etc 449 

XLJX.        Erection  of  Eads  and  Washington  Bridges 469 

L.  Erection  of  Niagara  Railway  and  Highway  Arches 471 

LI.  Erection  of  the  Kaiser  Wilhelm  Bridge 473 

LII.  Erection  of  the  Garabit  Viaduct  and  the  Bonn  and  Diisseldorf  Bridges 475 

LIIl.  Erection  of  the  Fairmount  Park  and  the  Rochester  Driving  Park  Bridges. .  477 

LIV.  Erection  of  the  Lake  Street  and  Panther  Hollow  Bridges 477 

LV.  Erection  of  the  Kornhaus  Bridge 477 

LVI.  Plans,  etc.,  of  Main  Steam  Piping  in  Power-Stations 485 

LVIl.         General  Diagram  of  Connections,  Port  Morris  Power-Station 487 

LVIII.        Plans  of  Port  Morris  Swit.h-House 489 

LIX.  Sections,  Port  Morris  Switch- House 491 

LX.  Wiring  Diagram,  Typical  Sub-Station 493 

LXI.  Diagram  of  Positive  Feeders,  N.  Y.  C.  &  H.  R.  R.  R.  Electrification 495 

LXII.         Cross-Section.  Removal  of  Grand  Central  Train-Shed 497 

LXin.       Traveler  for  Removal  of  Grand  Central  Train  Shed ' 499 

LXI V.        Views  Showing  Removal  of  Grand  Central  Train-Shed .  501 

LX V.         Appliances  used  in  Caissons  for  Foundations  of  Buildings 523 


418  CURVE   RESISTANCE  IN  WATER  PIPE  [Pai)crs. 

straight  line  composed  of  six  pipe  lengths  was  set  up  in  the  pipe  alley 
alongside  the  large  canal  of  the  Cornell  University  Hydraulic  Labora- 
tory. The  pipes  were  flange-connected.  The  total  length  of  6-in. 
pipe  was  122  ft.  The  loss  of  head  was  measured  in  a  length  of  99.33 
ft.,  a  length  of  20.04  ft.  being  allowed  up  stream  from  the  first 
piezometer  for  the  disturbances  due  to  entry  from  a  12-in.  header  to 
die  down.  The  piezometers  consisted  each  of  two  diametrically  op- 
])osito  holes  in  the  pipe  wall  into  which  i-in.  T-handle  cocks  were 
screwed.  A  three-way  connection  served  to  join  the  short  hoses  from 
these  cocks  and  the  single  long  hose  to  the  gauge.  A  water  differential 
gauge  was  used  to  measure  the  loss  of  head.  A  calibrated  concrete 
measuring  tank,  of  500  cu.  ft.  capacity,  received  the  discharged  water 
from  a  4-in.  pipe  with  a  4-in.  regulating  valve  through  which  the  6-in. 
line  discharged.  An  instantaneous  diverter  deflected  the  discharge  into 
the  tank  or  allowed  it  to  run  to  waste,  as  desired.  The  measurements 
lasted  from  5  to  10  min.,  the  time  being  accurately  taken.  Fig.  1. 
Plate  XL VI,  is  a  photograph  showing  the  pipe  line.  The  results  of 
these  measurements  are  given  in  Table  2,  where  data  from  later 
straight  pipe  experiments  are  given  also. 

After  this  the  pipes  were  marked,  disconnected,  and  transported 
to  the  bottom  of  the  Fall  Creek  Gorge,  in  Ithaca,  near  the  hydro- 
electric plant  of  Cornell  University.  Here  the  curve  experiments 
were  performed.  Fig.  2,  Plate  XL VI,  is  a  photograph  of  this  plant 
and  Fig.  1  is  a  plan  of  the  pipe  line. 

The  pipes  composing  the  experimental  portion  of  the  line  were  the 
same,  and  were  set  up  in  the  same  order  as  when  tested  as  a  straight 
pipe  line.  The  curves  were  placed  between  Pipes  Nos.  1  and  2,  the 
two  up-stream  pipes  in  the  experiments  described  above,  and  as  shown 
by  Fig.  1.  The  down-stream  tangent  portion  thus  consisted  of  Pipes 
Nos.  2,  3,  4,  and  5.  Pipe  No.  6  was  not  used  for  the  curve  experiments 
because  this  would  have  brought  the  end  of  the  pipe  line  out  into  the 
creek. 

At  the  discharge  end  of  Pipe  No.  5  there  was  attached  a  brass 
nozzle  for  measuring  the  velocity  in  the  pipe  line.  This  nozzle  is 
shown  by  Fig.  ?>,  Plate  XLVI.  It  had  been  calibrated  previously  by 
tank  measurements  in  the  Hydraulic  Laboratory.  The  average  dis- 
charge coefllcient  from  40  experiments,  with  pressure  heads  at  the 
base  of  the  nozzV^  ranging  from  1.G49  to  50.208  ft.,  was  found  to  be 
0.9S8. 


PLATE   XLVl. 

PAPERS,  AM.  SOC.  C.   E. 

MAY,   1908. 

SCHODER   ON 

CURVE   RESISTANCE   IN   WATER   PIPES. 


Fig.  1.— Six-Inth   Cipk  I.ink  <  ai-  iiii;  Kight) 
AS  Set  Up  for  First  Straight- 
Pipe  Experiments. 


Fig.  3.— Cornell  University  Hydro- 
Electric  Power  Plant,  and  6-In. 
Pipe  Line  for  Curve  Experiments. 


Fig.  3.— Nozzle  at  End  of  6-In.  Pipe  Line. 


Fig.  4.— Six-Inch,  90"  Curves,  Nos.  1  to  12. 


Papers.] 


CURVE  RESISTANCE  IN  WATER  PIPE 


419 


The  lengths  of  the  pipes,  from  face  to  face  of  flanges,  are  given  ixi 
Fig.  1;  the  following  are  the  inside  diameters — the  means  of  four 
measurements,  two  at  each  end: 


Pipe  No.  1. 
6.106  in. 


No.  2. 
6.086  in. 


No.  3. 
6.102  in. 


No.  4. 
6.078  in. 


No.  5. 
6.072  in. 


No.  6. 
6.083  in. 


Pipe  No.  5 

— 22;i»:!0-ft7- 
Down-st  i-L-am 
Piezometer 


Pipe  No.  t 
—22.218  ft- 


Pipe  No.  3 
— 19::J50ft 


Pipe  No.  2 
-20.GG0-ft— 


PLAN  OF 

6-INCH  WROUGHT- IRON 
PIPELINE, 

AS  USED  IN 

CURVE  EXPERIMENTS 


Up-stream- 
Piezometer 


oS5 

CO 

3  .5 


Fig.  1. 

The  Curves. — The  curves  used  were  made  to  order,  except  that 
Nos.  10  and  12,  respectively,  were  standard  "long  sweep"  and  "short 
turn"  6-in.,  cast-iron,  flanged,  90°  elbows.  Curves  Nos.  1  to  6,  in- 
clusive, were  bent  from  6-in.  wroiight-iron  pipe.  Curves  Nos.  7  to  12, 
inclusive,  were  of  cast  iron.  All  were  90°  curves;  and  all  were  left 
uncoated. 

The  wrought-iron  pipe  curves  had  al)out  6  in.  of  straight  pipe  at 
each  end.  The  ends  had  been  threaded  to  receive  standard  flanges. 
The  cast-iron  curves  were  flanged,  faced,  and  drilled  complete,  ready 
for  setting  up.  Fig.  4,  Plate  XL VI,  is  a  photograph  of  the  curves 
stacked  against  the  power-plant  wall. 


4^0 


CURVE   RESISTANCE  IN   WATER  PIPE 

TABLE  1. — Dimensions  of  Curves. 


[Papers. 


01 

0) 

4) 

a 

am 
a  S 

0) 

1 

0) 

a  « 
o  a_^ 

Material. 

a 

t;  1. 0) 

3 

■•^ 

CQ 

fl-S  a 

■a 

OS'S 

'^ 

oq 

+ 

tf 

M 

-r; 

Wrought  iron. 

10.00 

30 

10.54 

10.52 

21.06 

16.77 

7.50 

15 

8.04 

8.02 

16.06 

12.84 

11             ti 

5.00 

10 

5.59 

5.57 

11.16 

9.01 

"             " 

4.00 

8 

4.54 

4.. 52 

9.06 

7.34 

"             " 

3.00 

6 

3.60 

3.58 

7.18 

5.89 

2..'^0 

5 

3.05 

3.10 

6.15 

5.08 

Cast  iron. 

2.00 

4 

2.25 

2.25 

4.50 

3.64 

"       " 

l.CO 

3 

1.75 

1.75 

3.50 

2.86 

1.08 

2.16 

1.50 

1.50 

3.00 

2.54 

I.             n 

0.95 

1.90 

1.08 

1.08 

2.16 

1.75 

"       " 

0.88 

1.76 

2.00 

2.00 

4.00 

3.62 

0.67 

1.34 

0.67 

0.67 

1.34 

1.05 

cs  S  ce 


6.09in 

6.18 

6.16 

6.11 

6.11 

6.09 

5.91 

5.95 

5.91 

5.91 

5.95 

5.93 


The  dimensions  of  the  cvirves  are  given  in  Table  1,  in  which  ref- 
erence is  made  to  the  dimensions  indicated  in  Fig.  2.     The  dimensions 


in  Table  1  are  for  the  curves  as  placed  in  the  pipe  line,  the  vvrought- 
iron  bends  having  screw  flanges  attached.  The  distances  from  the 
faces  of  the  flanges  to  the  ends  of  the  wrought-iron  curves  were  as 
follows:  Curves  Nos.  1,  2,  3,  and  4,  up-stream  end,  i  in.,  down-stream 
end,  ^1  in.;  Curve  No,  5,  up-stream  end,  i  in.,  down-stream  end,  |  in; 
Curve  No.  6,  up-stream  end,  1  in.,  down-stream  end,  f  in.  The  inside 
diameters  of  Curve  No.  6  were  measured  also  at  points  22^°,  45°  and 
G7^°  from  the  up-stream  end.  At  these  points,  i-in.  taps  for  the  in- 
sertion of  a  Pitot  tube  had  been  made  on  both  the  vertical  and  hori- 
zontal diameters.  The  measurements  were  as  follows:  at  22i°,  5.92 
and  6.09  in.;  at  45°,  6.08  and  6.08  in.;  at  67^,  5.95  and  6.15  in. 


Papers.]  CURVE  KEiSlSTANCE  IN   WATER  PIPE  421 

Curves  Nos.  9  and  11  were  made  in  error,  the  foundry  making 
curves  of  quite  short  radius,  with  straight  portions  at  each  end,  so 
that  the  dimensions  from  the  center  to  the  face  were  18  and  24  in., 
respectively,  where  the  writer's  order  called  for  curves  with  the  same 
radii  as  the  center-to-face  dimensions.  Curves  Nos.  7  and  8  were  made 
later,  to  correct  this  error. 

The  experimental  pipe  line  was  arranged  with  the  idea  of  keeping 
all  conditions  the  same  except  for  the  introduction  of  the  several 
curves.  Consequently,  the  flange  joints  in  the  portion  down  stream 
from  the  curves  were  not  disturbed  throughout  the  experiments.  The 
entire  length  of  85  ft.  of  6-in.  wrought-iron  pipe,  together  with  the 
nozzle,  was  shifted  bodily  when  a  new  curve  was  placed  in  the  line. 
This  was  rather  vigorous  exercise  for  two  men,  but  it  was  accomplished 
by  a  judicious  tilting  of  the  wooden  horses  supporting  that  part  of 
the  pipe  line.  The  portion  up  stream  from  the  curves  was  left  un- 
changed during  the  experiments. 

The  same  up-stream  piezometer  was  used  for  the  curve  experi- 
ments as  for  the  first  straight-pipe  experiments.  The  down-stream 
piezometer  was  a  new  one,  of  the  same  type,  placed  1.05  ft.  up  stream 
from  the  down-stream  end  of  Pipe  No.  5  which  adjoined  the  nozzle. 
A  similar  intermediate  piezometer  was  placed  2.00  ft.  up  stream  from 
the  down-stream  end  of  Pipe  No.  3. 

At  first  it  was  assumed  that  some  effect  of  the  curves  might  ex- 
tend 100  or  more  diameters  down  stream  in  the  straight  pipe  beyond.* 
The  losses  of  head  in  all  the  curve  experiments  were  measured  be- 
tween -the  piezometer  just  up  stream  from  the  curves  and  the  piezometer 
just  up  stream  from  the  nozzle,  distant  168  diameters  down  stream 
from  the  curves.  These  two  piezometers  were  connected  to  the  two 
branches  of  a  differential  U-tube  mercury  gauge  by  lines  of  i-in.  or 
A-in.  three-ply  rubber  tubing.  The  nozzle  piezometer  was  connected  to 
one  branch  of  an  8-ft.  mercury  U-tube  gauge,  the  other  branch  of 
which  was  open  to  the  atmosphere.  These  gauges  were  provided  with 
blow-off  cocks  for  the  removal  of  air  from  the  gauge  and  connections. 

In  experimenting,  the  6-in.  valve  was  first  opened  wide  to  establish 
a  swift  flow  through  the  pipe  line.  The  gauge  hoses  connected  to  the 
piezometers  were  allowed  to  run  for  a  while  before  connecting  them 
to  the  gauges,  being  pinched  near  the  gauge  end  while  connecting  up. 

*  Transactions.  Am.  Soc.  C.  E.,  Vol.  XL VII,  1902,  p.  302. 


423  CURVE  RESISTANCE  IN  WATER  PIPE  [Papers 

Finally,  the  pet-cocks  were  opened,  and  all  traces  of  air  were  blown 
oft".  Then  the  gauge  readings  were  recorded  and  checked.  The  6-in. 
valve  was  then  closed  a  little,  and  the  readings  were  taken  again  as 
soon  as  the  flow  had  become  settled.  In  this  way  ten  or  twelve  runs 
were  made,  the  last  being  generally  a  repetition  of  the  first  run  with 
the  valve  wide  open.  If  there  was  time,  a  new  curve  was  then  placed 
in  the  line.  Otherwise,  the  change  was  deferred  until  another  dny, 
in  which  case  one  or  more  check  runs  were  made  at  the  high  velocities 
before  removing  the  curve  and  substituting  another.  For  each  set-up, 
the  level  of  the  center  of  the  nozzle  with  reference  to  the  nozzle 
mercury  gauge  scale  was  determined.  The  alignment  of  the  down- 
stream tangent  was  corrected  after  each  change  of  curves. 

When  the  work  on  the  twelve  curves  had  been  finished,  Curve  No. 
G  was  replaced  in  the  line  to  find  whether  or  not  any  appreciable 
change  in  the  condition  of  the  pipes  had  occurred  during  the  experi- 
ments. Then  a  number  of  Pitol-tube  studies  were  made  to  determine 
the  conditions  of  flow  at  the  piezometers  and  in  the  curve.  Later,  a 
6-in.  screw  elbow,  of  the  ordinary  steam-fitting  type,  was  placed  in 
the  line,  and  a  series  of  measurements  made  on  it  as  for  the  other 
curves.  Finally,  the  loss  of  head  was  measured  in  the  portion  (46.10 
ft.  long)  of  the  experimental  section  farthest  down  stream,  Pitot  tube 
traverses  having  shown  normal  flow  in  the  pipe  38  ft.,  or  76  diameters, 
down  stream  from  the  curves.  Hence,  this  portion  represented  straight 
pipe  unaffected  by  curvature  effects. 

The  results  of  these  loss-of-head  measurements  are  given  in  Table 
2.  For  simplicity,  there  is  given  only  the  mean  velocity  in  the  pipe 
line  as  deduced  from  the  nozzle  mercury  pressure  gauge  indications. 
The  differential  mercury  gauge  differences  are  taken  directly  from  the 
checked  subtractions  in  the  field  notebook.  These  differences,  multi- 
plied by  12.57,  would  give  the  loss  of  head,  in  feet  of  water,  the 
specific  gravity  of  mercury  being  taken  as  13.57. 

To  illustrate  the  calculations  involved  in  obtaining  the  velocity 
ill  the  pipe  line  from  the  observations  in  the  mercury  pressure  gauge 
the  following  is  a  sample: 

In  Experiment  No.  1,  Curve  No.  1,  October  10th,  1907,  the  gauge 
readings  were:  left  0.270,  right  7.900.  The  difference  is  7.630  ft. 
The  center  of  the  nozzle-tip  level  was  at  1.81  on  the  gauge  scale. 
ITcnce  the  top  of  the  left  mercury  column  was  1.54  ft.  below  the 
ccnlcr  of  the  nozzle  tip.     The  pressure  head  at  the  base  of  the  nozzle, 


PLATE  XLVII. 

PAPERS.  AM.  SOC.  C.  E. 

MAY,  1903. 

8CH0DER  ON 

CURVE  RESISTANCE  IN  WATER  PIPES. 


6-INCH   PIPE,  90°CURVE  EXPERIMENTS. 
Logarithmic  Plotting  of  Observed  Difterentlnl  Mercury  Gauge  Differences,  with 
Mean  Velocities  In  the  Pipe  lino,  showing  the  Losses  of  Head  in  the  E-xpcrimental 
Length  as  varied  only  by  the  introduction  of  the  several  Curves.  Also  the  corre- 
sponding Plotting  for  Straight  Pipe  without  Curvature  Effects.  _ 


Differential  Mercury  Gauge  Difference,  Feet 


Velocity,  in  Feet  per  Second 


I 


♦ 

Foldout 
Here 

♦  ♦ 

♦ 


II 


tC.li)  Ft.  o£ 
;raight  Pipe 
out  curvature 
etfects. 


Papers.]  CURVE  RESISTANCE  IN  WATER  PIPE  4'23 

therefore,  was  7.G30  X  13.57  —  1.54  =  102.1  ft.  The  diameter  of  the 
nozzle  tip  was  2.738  in.,  and  the  diameter  of  the  nozzle  base  was  6.128 
in.,  the  ratio  of  areas  being  1  :  5.    Theoretically,  the  nozzle-tip  velocity 


is    C  \j  /I  \2  ■/  2  gk=  C'X  S.l'.»  s/'h,  where  //,  is  the  pressure  lu^ad 

at  the  nozzle  l)ase.  As  stated  above,  tlie  coettieient  C,  for  this  nozzle 
had  been  found  io  be  0.988  by  experiment.  The  mean  diameter  of  the 
pipe  line  is  6.084  in.    Therefore  the  mean  velocity  in  the  experimental 

portion  of  the  pipe  line  was:  V  =  0.988  X  8.19  X     (g4li)  '^   ^^^  " 

1.639  v'/t.  This  gives,  for  this  experiment,  V  =  1.639  X  \/l02.T  = 
16.56  ft.  per  sec. 

The  results  given  in  Table  2  were  plotted  logarithmically  on  10-in. 
base  paper.     Plate  XLVII  shows  these  plottings  assembled. 

Now,  in  order  to  study  the  effects  of  the  curves,  it  is  necessary  to 
select  a  uniform  basis  of  comparison.  Thus  the  experiments  were 
made  by  introducing  the  various  curves  between  two  fixed  lengths  of 
pipe.  For  this  case,  the  observed  losses  of  head  may  be  compared  with 
each  other  at  once  by  using  Table  2  or  Plate  XLVII ;  but  such  a  basis 
takes  no  account  of  the  comparative  lengths  of  the  several  pipe  lines, 
nor  does  it  consider  whether  or  not  the  same  two  points  are  joined  by 
the  combinations  of  straight  pipes  and  curves. 

The  writer  has  chosen  two  bases  of  comparison.  In  one,  all  cases 
are  reduced  to  equal  lengths  on  the  center  lines  of  straight  pipes  and 
curves.  In  the  other,  all  cases  are  reduced  to  what  they  would  be  if 
two  fixed  points  had  been  connected  by  two  straight  pipe  lines  with 
the  various  90°  curves  between  them. 

For  both  these  cases  it  is  necessary  to  know  the  loss  of  head  in  the 
straight  pipe  when  it  is  unaffected  by  curvature.  Two  series  of  ex- 
periments already  described  give  this  information. 

Turning  to  Plate  XLVII,  it  is  seen  that  the  results  of  the  first 
straight-pipe  experiments  do  not  agree  with  the  later  ones.  Each 
series  gives  a  good  straight  line,  but  the  two  lines  have  different  slopes 
and  represent  different  laws  of  flow.  The  equations  of  these  lines,  re- 
duced so  as  to  represent  loss  of  head  in  feet  of  water  per  foot  length 
of  pipe,  are 

For  the  llrst  series,   ff  =  0.000  OC!)  V^-^' 
For  the   later  series,  7/ =  0.000  5'.>S  F'--'* 


426 


CURVE  RESISTANCE  IN  WATER  PIPE 


[  Pa  pcis. 


TABLE  2   (Coniinued). 


Date, 
1907. 


Oct.  19. 


Oc-(.  21. 


Oct.  24. 


n^. 


g  btxu 

53  o3  a 
S  >;« 
jj  o  aj 

>    In   bi 

life 


Curve  No.  11. 


1.050 
0.S126 
().f)07 
0.320 
0.275 
0.241 
0.202 
0.120 
0.0.56 
0.032 
1.046 

1.052 
0.715 


16.59 
15.53 
12.40 
8.99 
8.41 
7.75 
7.07 
5.41 
3.78 
2.87 
16.56 

16.58 
13.60 


Curve  No.  12. 


1.046 
1.009 
0.818 
0.558 
0.320 
0.168 
0.072 
0.087 
0.612 
1.040 


16.62 

16.31 

14.65 

12.01 

8.92 

6.43 

4.21 

2.74 

12.61 

Ki.Ol 


Screw  Elbow. 


Sag 


Date, 

1907. 


0 

a 

OJ 

« 

x> 

i 

<o 

a. 

!z; 

Ed 

vg  I 


o  a; 
w 
■■9 


01  r- 

6JD-.S 


47^ 


Straight  Pipe  Experiments. 


\  Temperature  of  water: 
First  Series:  -.  68°  at  beginning, 
I  70"  at  end. 


43 


43 


Nov.  12.. 

1 

1.090 

16.55 

37 

2 

1.036 

16.15 

Nov.  18.. 

3 

1.098 

16.55 

33 

4 

1.053 

16.21 

0 

0.912 

15.02 

6 

0.682 

12.94 

7 

0.515 

11.20 

8 

0.331 

8.93 

9 

0.174 

6.40 

10 

0.071 

4.04 

11 

1.088 

16.55 

Sept.  4.. 

1 

1.901 

0.0701 

6.03 

2 

1.687 

0.0622 

5.67 

3 

1.449 

0.0535 

5.22 

4 

1.241 

0.0458 

4.80 

5 

1.000 

0.0369 

4.29 

fi 

0.881 

0.0307 

3.88 

7 

0.704 

0.0260 

3.55 

8 

0.473 

0.0175 

2.87 

9 

0.307 

0.0114 

2.27 

10 

0.160 

0.0059 

1.61 

Sci- 

,„,1  Series*'  * 'r*'"'P*"''*^'^i"'2  *?'' 
.iKiheiies  .  1     water, ;«°  fain-. 

&"™=' 

Nov  18.. 

1 

13.70        0.506 

16.56 

2 

13.19        0.487 

16.23 

3 

10.24         0.378 

14.29 

4 

6.83         0.2525 

11.56 

5 

4.40        0.1625 

9.26 

6 

2.45         0.0905 

6.93 

*  The 

velocities  for  the  second  series  are 

compute 

d  for  the  mean  diameter  of  the  ex- 

peruneu 
6.075  in. 

tal  section  (46.10  ft.  long)  which  was 

Papers.]  CURVE  RESISTANCE  IN  WATER  PIPE  427 

To  show  more  clearly  the  magnitude  of  this  difference,  the  follow 
ing  caleiilated  values  are  given: 

Loss  of  head,  in  feet  of  water  per  foot  of  length : 


T'=  3  ft. 

F  =  5  f  t. 

F  =  10  ft. 

I'=  10  ft, 

per  see. 

per  sec. 

per  sec. 

per  see. 

First  series 

o.oons 

0.0135 

0.049 

0.119 

Later   series 

0.0050 

0.0135 

0.052 

0.129 

Thus,  at  the  outset,  comes  the  question  which  must  arise  in  all 
experiments  of  this  kind:  What  was  the  law  of  flow  of  the  identical 
straight  pipe  used  with  the  curves  when  unaffected  by  curvature,  but 
otherwise  in  the  same  condition  ?  On  this  depends  the  calculation  of 
the  excess  loss  of  head  caused  by  the  curves. 

Thore  are  a  number  of  possible  causes  for  the  difference  shown 
above : 

1. — The  pipe  may  have  become  rougher  by  rusting  in  the  interval 
between  the  two  series  of  experiments; 

2. — Pipes  Nos.  4  and  5  together  may  have  had  different  hydraulic 
properties  from  Pipes  Nos.  2,  3,  4,  and  5  together; 

3. — The  different  temperatures  of  the  water  may  have  caused  a 
difference  in  loss  of  head. 

It  was  intended  originally  to  bring  the  straight  pipes  back  to  the 
hydraulic  laboratory  and  again  test  them  as  at  first  for  loss  of  head 
after  the  curve  experiments  had  been  finished,  but  the  lateness  of  the 
season  prevented  this. 

Later  in  this  paper  the  above  possible  causes  of  differences  will  be 
discussed  more  fully.  In  order  to  remove  any  qtiestion  as  to  mistaken 
judgment,  the  results  are  worked  up  in  both  ways. 

The  individual  observations,  directly,  are  not  used  in  the  final 
comparisons.  From  the  mean  lines  drawn  on  the  logarithmic  diagram, 
where  all  observed  values  for  each  curve  have  been  plotted,  the  gauge- 
difference  values  for  velocities  of  3,  5,  10,  and  16  ft.  per  sec.  have 
been  picked  off.     These  values  are  given  in  Table  3. 

All  the  cases  have  been  reduced  to  the  length  conditions  existing  in 
the  set-up  for  Curve  No.  1.  For  the  first  comparison,  the  observed  dif- 
ferential mercury-gauge  differences  for  the  other  curves  have  been 
increased  by  an  amount  corresponding  to  the  additional  length  of 
straight  pipe  necessary  to  give  the  same  length  on  the  center  line  as 


«"?     we-: 


%t  ,i  , 


Axil  •.»l>'.J}(!h'.Hj>i«»J 


f*^ 


•»   n  -die  -jmn 


■>   MS 

•t.Ki 
•I  *1 


'I  A  I;  1,1 


i^X*  TO  RFDfCT.  Alt. 


I 


Nt. 

of  ciirv*'. 

1.  r 

:j 

7  75 

4 

tt.42 
lO.KT 
11. «8 

~ 

13.12 

For  equal       To  connec-t 
9o  <rf  e«rre.        Imsttsoo         two  fixed 
,  crater  Bnes. 


8 

9 
10 
11 
12 
Screw  elbow. 


14.28  • 

13.01  " 

1.3.14  - 

15,71  '• 


points. 


17.56  ft. 

18.06  •' 
1H.90  '■ 
17.06  " 
19.72    • 

20. n«    •• 


.1 


gives  the  len^  of  straight  pipe  to  be  added  for  eacli 
e  two  cases  to  rluce  all  to  the  conditions  of  Curve  No.  1. 
<>i  liead  per  footength  of  straight  pipe  is  given  in  Table  b. 
ables  r>  and  4,  he  individual  corrections  are  calculated, 
ults  are  given  in  Tae  C. 


m^ 


Papers.] 


CURVE  RESISTANCE  IN  \  ATKR   VIPE 

TABLE  f 


4:2\) 


Velocity, 

in  feel 

per  second. 

From  thk  Firsst  Experiments  :         From  the  Second  Experiments  : 

Mercury  difFerential-         -  ^    .       Mercury  differential- 
gauge  differences             wrjop            gauge  differences, 
in  feet.                      water.                   j^^  ^^^ 

Feet  of 
water. 

3 
5 
10 
16 

0.000M2                         0.00518                    0.000399 
0.001071                   :         0.01346                     0.001075 
0.00;»03                          0.0491                       0.00412 
0.00943                            0.1185                       0.01026 

0.00502 
0.0135 
0.0518 
0.129 

TABLE  U. — CoRRKCTioNs  TO  BE  Added  '1  Observed  Differential  Mer- 
cury-Gauge  DiFFERE.NCES   TO    ReDIT".    AlL   CaSES  FOR   COMPARISON 

WITH  Conditions  of  Curve  No.  1,  :  Feet. 


For  Equal  Lenuths  on  Center   Lines: 

1.) 

Connect  Two  Fixed  Points: 

No  of 

Velocity,  in  feet  per  second. 

No,f 
car\ 

Velocity,  in  feet  per  second. 

curve. 

3 

6 

10           16 

3 

5 

10 

16 

On  Basis  of  First  StraightPipe  Experiments. 


2 

0.0016 

0.004-2 

0.015 

0.087 

,-, 

0.0021 

0.0054 

0.020 

0.047 

8 

0.0082 

0.0OS3 

O.OMO 

0.078 

b 

0.0041 

0.0106 

0.039 

0.093 

4 

0.0089 

0.0101 

0.037 

0.089 

4 

0.0049 

0.0129 

0.017 

0.113 

5 

0.0M5 
0.0348 

0.0117 

0.042 

0.103 

5 

0.0057 

0.0149 

0.054 

0.131 

6 

0.0125 

0.046 

0.110 

f, 

0.0061 

0.0160 

0.058 

0.141 

7 

0.0054 

0.0141 

0.051 

0.124 

O.OOIW 

0.0177 

0.065 

0.156 

8 

0.0057 

0.0149 

0.054 

0.131 

0.0072 

0.01«8 

0.069 

0.166 

9 

0.0059 

0.01.52 

0.056 

0.134 

0.0074 

0.0194 

0.070 

0.170 

10 

0.0068 

0.0161 

0.059 

0.142 

1( 

0.0078 

0.0202 

0.074 

0.178 

11 

0.0054 

0.0141 

0.051 

0.1-24 

.11 

0.0070 

0.0183 

0.067 

0.161 

12 

0.0065 

0.0168 

0.061 

0.148 

12 

0.0081 

0.0211 

0.t77 

0.186 

Serf  1 
elbc  f 

0.0085 

0.0220 

0.080 

0.194 

On  Basis  of  Second  Straigf-Pipe  Experiments. 


2 

0.0010 

0.0042 

0.016 

0.040 

.' 

0.0020 

0.00.54 

0.021 

0.051 

3 

0.0031 

0.0083 

0.082 

0.080 

i. 

0.0040 

0.0106 

0.041 

0.102 

4 

0.0038 

0.0102 

0.038 

0.097 

0 

0.0048 

0.0129 

0.049 

0.123 

5 

0.0043 

0.0117 

0.045 

0.112 

I 

0.0055 

0.0149 

0.0.57 

0.142 

6 

0.0047 

0.0126 

0.048 

0.120 

f 

0.00.59 

0.0160 

0.061 

0.153 

7 

0.00.52 

0.0141 

0.054 

0.135 

t^ 

0.0066 

0.0178 

0.068 

0.170 

8 

0.0055 

0.0150 

0.0.57 

0.143 

i 

0.0070 

0.01H9 

0.072 

0.180 

e 

0  0057 

0.0153 

0.(i,'i9 

0.146 

f 

0.0072 

0.0194 

0.074 

0.185 

10 

0.0060 

0.0101 

0.002 

0.154 

U 

0.0075 

0.0203 

0.078 

0.194 

11 

0.0052 

0.0141 

0.054 

0.1.35 

11 

0.0068 

0.0183 

0.070 

0.175 

12 

0.0063 

0.0169 

0.065 

0.161 

l; 

0.0079 

0.0212 

0.081 

0.202 

Serf  ( 
elbef 

0.0082 

0.0221 

0.085 

0.211 

428 


CURVE  RESISTANCE  IN  WATER  PIPE 


[Papers. 


existed  between  the  two  piezometers  when  Curve  No.  1  was  in  the  pipe 
line.  For  the  second  comparison,  the  added  quantity  corresponds  to 
the  extra  length  of  pipe  required  to  make  tlie  sum  of  the  tangent 
distances  from  the  point  of  intersection  equal  to  that  for  Curve  No.  1. 
These  two  cases  may  be  called,  for  brevity,  respectively,  the  equal- 
lengths  and  the  two-fixed-points  cases. 

TABLE  3. — Differential  Mercury-Gauge  Differences  from  the 
Mean  Lines  Drawn  for  the  Plotted  Points  for  All  the  Ob- 
servations, IN  Feet. 


No.  of  curve. 

Velocity,  in  Feet  per  Second  : 

3 

5 

10 

16 

1 

0.0444 

0.1180 

0.443 

1.090 

o 

0.04U8 

0,1083 

0.410 

l.OIl 

3 

0.0400 

0.1062 

0.402 

0.991 

4 

0.0400 

0.1062 

0.402 

0.991 

5 

0.0395 

0.1046 

0.396 

0.978 

6 

0.0376 

0.1008 

0.386 

0.960 

7 

0.0381 

0.1013 

0.385 

0.950 

8 

0.0375 

0.1002 

0.382 

0.942 

9 

0.0375 

0.1002 

0..382 

0.942 

10 

0.0375 

0.1002 

0.382 

0.942 

11 

0.0382 

0.1030 

0.394 

0.981 

12 

0.0387 

0.1031 

0.392 

0.968 

Screw  elbow. 

0.0403 

0.1077 

0.413 

1.030 

TABLE  4. — Lengths  of  Straight  Pipe  to  be  Added  to  Reditce  All 
Cases  to  Conditions  of  Curve  No.  1. 


No.  of  curve. 

For  equal 
lengths  on 
center  lines. 

3.92  ft. 
7.75     " 
9.42     '■ 
10.87    " 
11.68    •' 
13.12    " 

To  connect 

two  fixed 

points. 

No.  of  curve. 

For  equal 
lengths  on 
center  lines. 

To  connect 

two  fixed 

points. 

2 
3 
4 
5 
6 
7 

5.00  ft. 

9.90  •' 
12.00   " 
13.88   " 
14.91    " 
16.56   " 

8 

9 
10 
11 
12 
Screw  elbow. 

13.90  ft. 
14.22  '• 
15.01   " 
13.14   " 
15.71    " 

17.56  ft. 
18.06  •' 
18.90  " 
17.06  " 
19.72  '• 
20.56  " 

Table  4  gives  the  length  of  straight  pipe  to  be  added  for  eacli 
curve  for  the  two  cases  to  reduce  all  to  the  conditions  of  Curve  No.  1. 

The  loss  of  head  per  foot  length  of  straight  pipe  is  given  in  Table  r>. 

TTsing  Tables  r>  and  4,  the  individual  corrections  are  calculated. 
The  results  are  given  in  Table  C. 


Papers.] 


CURVE  RESISTANCE  IN  WATER  PIPE 

TABLE  5. 


40:» 


Velocity, 

in  feel 

per  second. 

From  the  First  Experiments  : 

From  the  Second  Experiments  : 

Mercury  differential- 
gauge'  difference  , 
in  feet. 

Feet  of 
water. 

Mercury  differential- 
gauge  differences, 
in  feet. 

Feet  of 
water. 

3 

5 
10 
16 

0.000412 
0.001071 
0.003903 
0.00943 

0.00518 
0.01346 
0.0491 
0.1185 

0.000399 
0.001075 
0.00412 
0.01026 

0.00502 
0.0135 
0.0518 
0.129 

TABLE  G. — CoRRFX'TiONs  to  be  Added  to  Observed  Differential  Mer- 
cuRv-G.\uGE  Differences  to  Reduce  All  Cases  for  Comparison 
WITH  Conditions  of  Curve  No.  1,  in  Feet. 


For  Equal  Lengths  on  Center  Lines: 

To  Connect  Two  Fixed  Points: 

No  of 

Velocity,  in  feet  per  second. 

No.  of 
curve. 

Velocity,  in  feet  per  second. 

curve. 

3 

5 

10            16 

3 

5 

10 

16 

On  Basis  of  Eirst  Straight-Pipe  Experiments. 


2 

0.0016 

0.0043 

0.015 

0.037 

2 

0.0021 

0.0054 

0.030 

0.047 

3 

0.0033 

0.0083 

O.OriO 

0.073 

3 

0.0041 

0.0106 

0.039 

0.093 

4 

0.0039 

0.0101 

0.037 

0.089 

4 

0.0049 

0.0129 

0.017 

0.113 

5 

0-0045 

0.0117 

0.043 

0.103 

5 

0.00.57 

0.0149 

0.054 

0.131 

6 

0.0948 

0.0125 

0.046 

0.110 

6 

0.0061 

0.0160 

0.058 

0.141 

7 

0.0054 

0.0141 

0.051 

0.134 

7 

0.0088 

0.0177 

0.065 

0.156 

8 

0.0057 

0.0149 

0.054 

0.131 

8 

0.0072 

0.0188 

0.069 

0.166 

9 

0.0059 

0.0152 

0.056 

0.134 

9 

0.0074 

0.0194 

0.070 

0.170 

10 

0.0062 

0.0161 

0.059 

0.143 

10 

0.0078 

0.0202 

0.074 

0.178 

11 

0.0054 

0.0141 

0.051 

0.124 

11 

0.0070 

0.0183 

0.067 

0.161 

12 

0.0065 

0.0168 

0.061 

0.148 

12 

0.0081 

0.0211 

0.C77 

0.186 

Screw  1 
elbow  f 

0.0085 

0.0220 

0.080 

0.191 

On  Basis  of  Second  Straight-Pipe  Experiments. 


2 

0.0016 

0.004<J 

0.016 

0.040 

2 

0.0020 

0.0054 

0.021 

0.051 

3 

0.0031 

0.0083 

0.032 

0.080 

3 

0.0040 

0.0106 

0.041 

0.102 

4 

0.0038 

0.0102 

0.038 

0.097 

4 

0.0048 

0.0129 

0.049 

0.123 

5 

0.0043 

0.0117 

0.045 

0.112 

5 

0.0055 

0.0149 

0,057 

0.142 

6 

0.0047 

0.0136 

0.048 

0.120 

6 

0.0059 

0.0160 

0.061 

0.153 

7 

0.0052 

0.0141 

0.054 

0.135 

7 

0.0066 

0.0178 

0.068 

0.170 

8 

0.0055 

0.0150 

0.057 

0.148 

8 

0.0070 

0.0189 

0.072 

0.180 

9 

0  0057 

0.0153 

0.(1.59 

0.146 

9 

0.0072 

0.0194 

0.074 

0.185 

10 

0.0060 

0.0161 

0.062 

0.154 

10 

0.0075 

0.0203 

0.078 

0.194 

11 

0.0052 

0.0141 

0.054 

0.135 

11 

0.0068 

0.0183 

0.070 

0.175 

12 

0.0063 

0.0169 

0.065 

0.161 

12 

0.0079 

0.0212 

0.081 

0.202 

Screw  ( 
elbow  f 

0.0082 

0.0221 

0.085 

0.211 

432 


CURVE  RESISTANCE  IN  WATER  PIPE 


[Papers. 


TABLE  9. — Lengths  of  Straight  Pipe,  in  Feet^  to  give  Loss  of  Head 

EQUAL  to  the  ExCESS  LoSS  DuE  TO   THE   CuRVES.      On  THE  BaSIS  OF 

Second  Straight-Pipe  Experiments. 


For  Rqual  Lengths  on  Center  Lines: 

To  Connect  Two  Fixed 

Points: 

Veloci 

ty,  in  feK 

per  second. 

Velocity,  in  feet 

per  second. 

No.  of 

No.  of 
curve. 

curve. 

3 

5 

10 
4.4 

16 

3 

5 

10 

16 

1 

8.4 

6.7 

3.2 

1 

4.0 

2.4 

0.2 

—1.1 

2 

3.2 

1.6 

0.2 

-0.6 

2 

o.O 

—1.6 

—2.7 

—3.8 

3 

5.0 

3.5 

2.1 

1.4 

3 

3.0 

1.3 

0.2 

-0.8 

4 

6.8 

5.3 

3  9 

3.0 

4 

5.0 

3.4 

2.1 

1.2 

.5 

6.8 

5.1 

3.9 

3.2 

5 

5.6 

3.8 

2.7 

1.9 

6 

3.0 

2.5 

2.1 

2.3 

6 

1.8 

1.3 

1.2 

1.2 

7 

5.6 

4.3 

3.5 

2.7 

7 

4.8 

3.4 

2.7 

1.9 

8 

4.8 

4.1 

3.5 

2.7 

8 

4.2 

3.4 

2.9 

2.0 

9 

5.2 

4.4 

3.9 

3.0 

9 

4.8 

3.9 

3.5 

2.6 

1(1 

6.0 

5.1 

4.6 

3.8 

10 

5.6 

4.7 

4.4 

3.4 

11 

5.8 

5.8 

5.6 

5.7 

11 

5.6 

5.5 

5.4 

5.3 

13 

9.8 

8.6 

7.7 

7.0 

12 

9.6 

8.3 

7.5 

6.7 

Screw  ( 
elbow.  ) 

14.3 

13.4 

13.5 

13.6 

TABLE  10. — Excess  Losses  of  Head  Due  to  Curves,  Expressed  in 
Terms  of  Velocity  Heads.  On  Basis  of  Second  Straight-Pipe 
Experiments. 


For  Equal  Lengths  on  Center  Lines: 

To  Connect  Two  Fixed  Points: 

Velocity,  in  feet  per  second. 

Velocity,  in  feet  per  second. 

No.  of 

3 

5 

10 

16 

No.  of 
curve. 

t- 

S 

TO 

(2 

curve. 

5^ 

*% 

in 

a 

o 

,_ 

TO 

II 

II 

II 

II 

3 

5 

10 

16 

fc.^ 

tLJ^ 

tL  <^ 

^  ^g" 

0.14 

1 

0.30 

0.23 

0.15 

0.10 

1 

0.08 

0.01 

-0.04 

2 

0.11 

0.05 

0.01 

-0.02 

2 

0.00 

—0.05 

—0.09 

—0.13 

3 

0.18 

0.12 

0.07 

0.05 

3 

0.11 

0.05 

O.dl 

-0.03 

4 

0.24 

0.18 

0.13 

0.10 

4 

0.18 

0.12 

0.07 

0.04 

5 

0.24 

0.18 

0.13 

0.10 

o 

0.20 

0.13 

0.09 

0.06 

6 

n.ii 

0.08 

0.07 

0.07 

6 

0.06 

0.05 

0.04 

0.04 

7 

0.20 

0.15 

0.12 

0.09 

1 

0.17 

0.13 

0.09 

0.06 

8 

0.17 

0.14 

0.12 

0.09 

8 

0.15 

0.12 

0.10 

0.07 

9 

0.19 

0.15 

0.13 

0.10 

9 

0.1? 

0.14 

0.12 

0.08 

10 

0.21 

0.18 

0.15 

0.12 

10 

0.30 

0.16 

0.15 

0.11 

11 

0.21 

0.20 

0.19 

0.19 

11 

0.30 

0.19 

C.18 

0.17 

12 

0.35 

0.30 

0.26 

0.33 

12 

0.34 

0.39 

0.35 

0.22 

Screw  elbow. 

0.52 

0.47 

0.45 

0.44 

Papers.] 


CURVE   RESISTANCE  IN   WATER  PIPE 


433 


tendency  represents  the  trntli,  tlien  a  remarkable  dilemma  is  presented. 
As  the  curvatvire  of  a  pipe  becomes  less  and  less  the  external  condi- 
tions approach  nearer  to  those  of  straight  pipe.  The  natural  in- 
ference is  that  the  loss  of  head  also  approaches  straight-pipe  values, 
unless,  indeed,  it  be  argued  that  the  slightest  deflection  from  straight 
pipe   immediately   causes    a   considerable   excess   loss   of   head.*     For- 


5  G 

Radius  of  Ciirve.in  Feet 
Figs.  3  and  4. 

tunately,  there  exist  experimental  data  which  will  assist  in  the  con- 
sideration of  these  points.     These  will  be  given  presently. 

Fig.  G  shows  the  same  tendency  as  Fig.  5,  although  the  decrease  in 
loss  of  head  with  increasing  radius  of  curvature  is  less  decided  on  the 
basis  of  the  first  straight-pipe  experiments. 

♦Detroit  Curve  Experiments,  Transactions,  Am.  Soc.  C.  E.,  Vol.  XLVII,  1902.  Cod- 
elusion  J,  page  191,  and  pages  186-187. 


434 


CURVE  RESISTANCE  IN  WATER   PIPE 


[Papers. 


As  to  the  diflFerence  found  between  tlie  results  of  the  two  straight- 
pipe  series,  the  following  may  be  stated.  Conditions  of  temperature 
were  favorable  for  some  rusting  during  September,  after  the  first 
straight-pipe  series  and  before  the  fii*st  curve  experiments  were  made. 


0.9 

0.8 

^ 

K 

"o.r 

y 

'C 

I 

0.0 

p 

3 

0.5 

to 

c 

■p 

5) 

0.4 

o 

C3 

0.3 

o 

~ 

■;; 

0.2 

'~' 

o 

^ 

a 

0.1 

7 

Q) 

s 

J3 

-0.1 

3h 

O 

- 

-3 

-0.2 

^ 

^ 

i 

-0.3 

1. 

o 

_o 

-0.4 

01 

C 

-0.5 

►^ 

1:14 

o 

bO 

V       1  >^ 

\ 

90°CURVE  EXPERIMENTS, 

6.INCH   PIPE 

\ 

T 

Y    1 

f., 

\ 

._r^ 

1 

„  "t-|T,?*;^'< 

.5j.Mi;-M*i^ 

.^.^ 

fe^ 



r^ 

w 

( 

i — - 

ss^ 

_ 

^■^ 

\ 

^'^^ 

N 

\ 

"^ 

r'' 

y 

N. 

\, 

/■ 

y 

v 

\ 

y 

^ 

\ 

/" 

4  5  6 

Radius  of  Curve, in  feet 


Radius  of  Curve.in  feet 
Figs.  5  a.nd  fi. 

During  the  curve  experiments,  the  data  show  no  indication  of  increas- 
ing roughness,  or  of  any  effect  of  changes  in  temperatiire  of  the 
water.*    All  the  straight  pipes  had  been  used  in  a  steam-heating  main 

*This  is  remarliable.  It  has  been  observed,  for  smooth  brass  pipes  of  all  sizes  between 
,'n  in.  and  5  in.  in  diameter,  that  the  loss  of  head  is  increased  about  4"„  for  a  decrease  of 
temperature  of  the  water  of  10"  fahr.  Rougher  pipes,  such  as  Ka'vaiiis'ed  iron  and  wrought 
iron,  show  no  effect  due  to  temperature  changes. 


Papers.  ] 


CURVE   KKSTSTANCE  IN   WATER  RIPE 


435 


for  some  years,  and  all  seemed  to  have  a  uniform  internal  appearance. 
The  first  series  had  no  velocities  greater  than  6  ft.  per  sec,  while  the 
second  series  had  velocities  as  hi^h  as  in  the  curve  experiments.     A 


~  s 


bo  '5    a 
o  i:  '5 


s 

1 

90°CURVE  EXPERIMENTS, 

/ 

i 

fc 

6-INCH  PIPE 

1 

s 

^ 

^* 

} 

1 
/ 

'"""' 

■-"•^. 

(^ 

,, 

y 

I 

k 

,.'^ 

, 

1 
1 
1 

V  — 

-4.  •-^' 

./ 

/ 

f 

4 

e-X4\ 

11 

^ 

.% 
'T^^ 

~^* 

^i^^ 

"•'• 

■~-.« 

^ 

/ 

/•■ 

A 

r-^ 

il 

^-^ 

"■-^ 

"^o" 

/ 

./ 

-^ 

-^ 

? 

^^ 

-J 

^ 

• 

' 

,/ 

y 

/ 

+-^ 

^-*, 

^^ 

T 

/V 

"^ 

^. 

^- 

^y 

y^^ 

-Y 

4  5  6 

Radius  of  t'urvo, in  Foot 
Fig,  7. 


10 


4  5  6 

lladiiis  of  <-'urve,in  Feet 
Fig.  8 


separate  measurement  of  the  loss  of  head  in  Pipes  l^os.  2  and  3  when 
Tuiinfluenced  hy  curvature  was  not  made,  and  it  is  impossible  to  de- 
cide as  to  their  hydra idic  properties  as  compared  with  Pipes  Nos.  \ 
and  5. 


436 


CURVE  RESISTANCE  IN  AVATER   PIPE 


[Papers. 


It  is  thus  clear  tliat,  without  additional  evidence,  it  is  not  possible 
to  reject  one  of  the  straight-pipe  series  and  accept  the  other,  or  to  feel 
safe  in  using  average  values. 

In  relation  to  this  matter,  the  writer  desires  to  present  the  data 
from  measurements  on  an  8-in.  cast-iron  water  main.  This  nuiin  sup- 
plies raw  water  to  the  Cornell  University  Filtration  Plant.*  It  had 
Ixeu  laid  and  in  use  for  three  years  before  the  experiments  in  the  fall 
(if  1!K)0.  Before  laying,  the  inside  diameter  of  each  length  had  been 
calipered.  At  the  time  of  laying,  each  pipe  length  was  set  accurately 
to  line  and  grade  with  a  transit.     After  laying,  and  before  covering 

CORNELL  UNIVERSITY 

8-INCH   CAST-IRON   RAW  WATER 

PIPE    LINE  This  Inlcrincilalt  tangent  contains 

,.,,.,      ,       ..^^^  -^  ,       .         .    ,        Change  of  grade  made 

lift.plpe  lengths. ^^^        Change  of  grade  made  b;       .  ,h    i,     d 


Change  of  graJe  made  b; 
five  equal  deflect; 
Change  of  grade  made  bj     conBecutire  pipe  joints 


O.^iC'lU'iii 


PROFILE 

LengthB  betnoeti  plczuiiiolers  on 

centur  line.  Feet 

B-f    HO.SU         I -J     106.06 


81.03 
82.82 
1IJ'J.07 


L-M 


i.28 


Mean  diameter^  IncbeB, 
£  -  F    7.999        /  -J    8.000 
F-a     8.O30       J-K    8.022, 
a-U    8.008       K-L    not  Measured 


900- 
8'JO 
880 

DiameterB,  Vertical,  at  .  -g^Q  - 

piezometer  tape.  Inches 

f    7M         /     7.98  M     7.S3-8(iO- 

f!   7.98        J    7.97 

a    8.0fi        K     8.05  850- 

fl    8.02        X     7.94 

810 

830 


^.OOii 


L    M 


a    II 


.l.tl.ctlo 
tioual  |.ol 


FiQ.  9. 

the  pipe,  the  piezometer  holes  were  drilled  and  tajiped,  and  the  diam- 
eters measvired  at  these  points;  the  ;i-in.  brass  pi<'/.onieter  cocks  were 
inserted  so  as  not  to  project  inside,  and  the  lengths  between  the 
piezometers  were  measure<l  and  checked  in  the  ditcli. 

The  plan  antl  profile  of  the  pipe  line  are  shown  by  Fig.  9.  There 
are  eight  experimental  sections,  four  of  which  contain  deflections,  and 
four  of  which  are  straight,  preceded  by  considerable  lengths  of  straight 
pipe. 

The  piezometer  taps  were  placed  on  top  of  the  pipe  and  1  ft.  up 
stream  from  the  joints,  except  that  those  before  deflections  were  placed 
2  ft.  up  stream  from  the  joint  where  the  first  deflection  occurred. 

The  flow  in  the  pipe  line  was  measured  at  the  filter  plant  by  a 


*Designerl  by  G.  S.  Williams,  M.  Am.  Soc.  C.  E.    See  The  EiKjinecn'iin  Record,  April  Otli,  ItVM. 


Papers.] 


CUHVK   KKSISTANCK   IN    WATKR    I'll'K 


437 


10 

" 

/ 

8 

/ 

¥ 

'i 
6 

5 

4 

3.5 

3 

•^.5 

ri 

oi 

• 

ect 

on 

E-F 

J 

o 

+ 

I-J 
K- 

L 

f 

t 

0 

f 

r 

T 

1.5 
1 

^ 

-1 

i 

c 

>' 

f 

; 

/ 

/ 

^ 

0.9 
0.8 
0.7 
O.C 

0.5 
0.4 

r 

/ 

/ 

8-INCH   CAST-IRON   PIPE 

7 

to 

3ARITHWIIC  PLOTTING 
FOR 

^ 

' 

s 

Sec 

TRAIGHT  SECTIONS 
;  Fig.9  for  Dimensions 

/^ 

f 

0.3 

( 
/ 

/ 

/ 

0.5     0.6    0.7  0.8  0.9   1  1.5  2  2.5 

Velocity,  in  Feet  per  Second 
Fig.  10, 


438  CURVE   RESISTANCE  IN   WATEK    I'll'i:  [I'apcis. 

Venturi  meter  which  had  been  accurately  calibrated  in  place  by 
volumetric  measurements.  A  differential  water  gauge  was  used  with 
the  meter.  The  losses  of  head  were  measured  with  a  portable  differ- 
ential water  gauge  mounted  on  a  tripod.  The  gauge  was  set  up  on 
the  ground  midway  between  two  piezometer  wells,  and  pressure  con- 
nections were  made  with  small  three-ply  rul)ber  hose  after  thoroughly 
blowing  off  to  remove  all  air. 

The  flow  was  controlled  by  a"  valve  at  the  Alter  plant.  It  was  i)OS- 
sible  to  shut  this  valve  down  entirely,  and  thus  get  no-flow  conditions 
and  a  check  on  the  gauge  readings.  The  electric  motor-driven  two- 
stage  centrifugal  pumps  at  the  lower  end  of  the  pipe  line  allowed  this 
procedure  without  any  trouble. 

After  changing  the  valve  setting  and  allowing  the  flow  to  become 
steady,  simultaneous  readings  were  taken  on  the  meter  and  the  loss- 
of-head  gauges.  The  results  are  shown  graphically  in  Figs.  10  and  11. 
For  the  straight  sections  the  observed  losses  of  head  have  been  reduced 
uniformly  to  loss  per  1 000  ft.  It  will  be  seen  on  Fig.  10  that  the 
straight  sections  differ  among  themselves.  Thus  Section  E-F  has  less 
and  Section  K-L  has  greater  loss  of  head  than  the  average  of  the  four 
sections;  in  fact,  Section  K-L  has  about  15%  greater  loss  of  head 
than  Section  E-F.  The  equation  of  the  mean  line  for  the  four  sec- 
tions of  straight  pipe  is 

H  =:0.58G  T^^-^^ 
where  H  is  the  loss  of  head,  in  feet  per  1  000  ft.,  and  V  is  the  velocity, 
in  feet  per  second.     (The  corresponding  values  of  0,  in  T^  ^  C  sj  R  8, 
are:  at  1  ft.  per  sec,  101;  at  4  ft.  per  sec,  107.) 

Fig.  11  shows  the  difference  between  this  average  law  of  flow  for 
the  straight  portions  and  the  hydraulics  of  the  sections  containing  de- 
flections. It  is  remarkable  that  Section  F-G  with  a  single  3.8°  de- 
flection. Section  H-I  with  a  curve  composed  of  five  3.17°  deflections, 
and  Section  J-K  with  a  reverse  curve  composed  of  one  curve  with 
five  2.18°  deflections  and  another  with  four  2.81°  deflections,  all  show, 
on  the  whole,  less  loss  of  head  than  the  average  of  equal  lengths  of 
straight  pipe.  Section  L-M,  with  a  short-radius  bend  giving  a  deflec- 
tion of  12°  56',  is  the  only  one  showing  a  greater  loss  of  head. 

The  writer  does  not  argue  from  this  that  such  easy  curves  or  de- 
flections are  more  favorable  fm'   (lie  flow  of  water  than  straight  pipe. 


Papers.] 


ciMai;  i;i:s  I  STANCE  ix  watku  pipe 


439 


1 

0.9 
0.8 
0.7 
00 

/ 

f 

1    1    1  1  1  1  M  M  1  !  i 

•/ 

1 

/ 

8-lNCH  CAST-IRON   PIPE 

LOGARITHMIC  PLOTTING 

FOR  SECTIONS 

CONTAINING  DEFLECTIONS 

p  i 

f f 

f 

>/ 

J. 

/ 

/ 

11 

0.5 
0.45 
0.4 
0.35 

■»    0.3 
a) 

"cS 

F 

J/ 

4) 

/ 

n 

tA> 

1 

/ 

( 

/ 

/ 

/% 

8) 

/ 

( 

1 

t 

c 

/ 

r 

/ 

/ 

o/j 

>    0.S5 

o 

fa 

-o    0.15 

<a 
K 

0 
en 
» 

3    0.1 
%    0.09 
1    O.OS 
1    0.07 
0.06 
0.05 

0.04 
0.03 

on-" 

i 

1 

0 

y 

/ 

f 

• 

c 

V 

I 

/ 

f 

f 
j 

/ 

/ 

■ 

< 
/ 

V 

) 

/ 

/ 

/ 
/ 

/ 

/ 

i 

(^ 

/ 
/ 

) 

i 

• 

/ 

/ 

3 

f\ 

\f 

J 

/ 

/ 

/ 

/ 

/ 

/(t 

^ 

/ 

/ 

/ 

/ 

'1 

/ 

/ 

h 

V 

The  lines  are  not  drawn  to  fit  the  plotted  points, 
but  represent  the  eonditions  in  the  straight 
portions  of  the  pipe  line  for  lengths  equal  to 
the  lengths  of  the  sections  with  detlections. 

Hence  this  plotting  shows  graphically  the 
excess  or  deficiency  in  loss  of  head  due  to  the 
deflections. 

See  Fig.9  for  Plan  and  Profile  of  Pipe  Line 
with  dimensions. 

/ 

/ 

O 

7 

"// 

k) 

^ 

7 

0.5     0.6    O.T  0.8  0.9  1 


1.5  3        3.5       3 

Velocity,  in  Feet  per  Second 
Fig. 11. 


440  CURVE  RESISTANCE  IN  WATER  PIPE  [Papers. 

but  he  does  see  the  indication  that  any  difference  is  very  small  and 
may  be  less  than  the  difference  between  two  straight  sections  in  the 
same  pipe  line,  as  occurs  in  the  case  above  recorded. 

The  6-in.  wrought-iron  pipe  experiments  also  give  some  informa- 
tion on  the  question  of  the  effect  of  slight  deflections  in  otherwise 
straight  pipe.  The  first  series,  of  October  10th,  1907,  was  made  with 
a  decidedly  zigzag  appearance  of  the  down-stream  tangent,  that  is, 
the  joints  were  not  in  a  straight  line,  although  the  individual  pipe 
k'ngths  themselves  were  straight.  On  October  12th  the  series  was  re- 
peated, but  with  the  down-stream  tangent  carefully  aligned.  No  dif- 
ference in  results  is  noticeable. 

Viewed  in  the  light  of  the  foregoing,  it  is  easy  to  decide  that  the 
first  6-in.  straight-pipe  experiments  do  not  apply  to  the  later  curve 
experiments  because  the  deduced  excess  loss  of  head  docs  not  continue 
to  approach  zero  .for  the  long  easy  curves.  In  this  respect,  the  differ- 
ence between  Figs.  3  and  4  is  noteworthy. 

Now,  all  of  this  contradicts  the  findings  of  Messrs.  Williams,  Hub- 
bell,  and  Fenkell  in  the  Detroit  Experiments.  The  writer  cannot 
imagine  that  radically  different'  laws  apply  to  the  cases  investigated 
by  him  and  by  these  experimenters. 

One  difference  in  conditions  is  to  be  noted,  however.  The  long- 
radius  curves  in  the  30-in.  Detroit  main  were  made  up  of  several 
pieces,  while  the  writer's  6-in.  curves  were  all  one-piece  bends.  As  to 
the  probable  small  effect  of  the  joints,  the  writer's  8-in.  pipe  experi- 
ments, with  small  deflections,  give  some  idea;  but  there  are  other  pos- 
sible causes  for  the  divergence  of  the  findings.  The  smallness  of  the 
measured  losses,  with  the  comparatively  low  velocities  available  in  the 
Detroit  30-in.  main,  would  tend  to  magnify  excess  losses  due  to  other 
effects  than  curvature.  Thus,  in  Figs.  3  and  5  it  will  be  seen  that  a 
very  different  appeal  to  the  eye  is  given  by  the  line  for  a  velocity  of 
16  ft.  per  sec.  than  by  the  line  for  a  velocity  of  3  or  even  5  ft.  per  sec. 

There  remains,  also,  for  the  Detroit  Experiments,  the  possibility 
of  relatively  large  errors  due  to  several  causes.  These  errors  were 
considered  by  Messrs.  Williams,  Hubbell,  and  Fenkell  in  their  closing 
discussion,  and  a  table  was  presented*  in  which  corrections — as  large 
fs  50%  in  one  case — were  made  to  the  results  given  in  the  main  part 
of  the  paper.  These  corrections  materially  alter  the  appearance  of 
the  remarkable  Fig.  90  of  the  paper. 

*Table  No.  89,  page  360,  Transactions,  Am  Soc.  C.  E.,  Vol.  XLVII,  1902. 


Papers.]  CURVE  RESISTANCE  IN  WATER  PIPE  441 

If,  now,  ill  the  Detroit  Experiments,  to  the  causes  for  uicorrect 
deductions  above  mentioned  there  be  added  the  effect  of  using  for  com- 
parison the  results  of  experiments  on  short  sections  of  straight  pipe 
that  might  have  had  quite  different  hydraulic  properties  from  the 
straight  pipe  in  the  curve  section,  it  is  easy  to  see  that  the  combination 
of  circumstances  may  have  led  to  conclusions  not  at  all  general  in 
their  applicability,  and  perhaps  even  wrong  for  the  case  iji  hand;  but, 
as  to  this  matter,  the  writer  is  quite  content  with  suggesting  the 
salient  arguments. 

After  all,  however,  the  engineer  will  be  particularly  interested  in 
the  magnitude  of  the  excess  losses  of  head  due  to  curves.  Are  they 
seriously  large  in  an  extreme  case? 

Figs.  7  and  8  show  that  the  excess  loss  of  head  for  the  shortest  90'^ 
curve  is  equal  to  the  loss  in  7  to  10  ft.  (or  14  to  20  diameters)  length  of 
straight  pipe.  When  the  radius  is  2^  diameters  the  excess  loss  of 
head  is  equal  to  the  loss  in  5  to  10  diameters  length  of  straight  pipe. 
In  this  relation  Fig.  12  is  interesting.  It  is  seen  that,  for  the  smooth 
brass  180°  curves,  the  excess  loss  of  head  is  rather  less  than  the  loss  of 
head     in     7     diameters     length    of     straight     pipe.     The     expression, 

045  =  n~>  seems  to  give  a  fair  average  value  for  the  range  of  these 
brass-curve  experiments. 

Table  9  shows  that  the  6-in.  screw  elbow  gives  an  excess  loss  of 
head  equal  to  the  loss  in  about  27  diameters  length  of  straight  wrought- 
iron  pipe.  The  writer  has  also  the  record  of  some  accurate  measure- 
ments on  the  loss  due  to  3-in.  and  4-in.  screw  elbows,  from  which  it 
appears  that  the  losses  are  equal,  respectively,  to  the  losses  in  25  and  27 
diameters  length  of  straight  wrought-iron  pipe. 

When,  therefore,  the  Detroit  Experiments,  after  thorough  revision, 
indicate  for  a  long  easy  curve  in  30-in.  pipe  an  excess  loss  of  head 
equal  to  the  loss  in  50  diameters  length  of  straight  pipe,  not  only  does 
tlie  loss  seem  to  be  too  large  when  compared  with  the  loss  in  a  screw 
elbow  where  sudden  enlargement  and  contraction  are  present  in  addi- 
tion to  extremely  short-turn  curvature  effects,  but  the  whole  trend  of 
the  results  is  directly  the  opposite  of  what  is  shown  by  the  writer's 
experiments  on  6  and  8-in.  pipes. 

Now,  it  may  be  that  Nature  changes  her  methods  somewhere  be- 
tween pipes   of  8   and   30   in.   in   diameter,   as   regards   the  effects  of 


443 


CURVE  RESISTANCE  IN   WATER   PIPE 


[Papers. 


curvature.  It  must  be  so,  if  both  the  Detroit  Experiments  and  those 
of  the  writer  have  been  interpreted  correctly.  The  evidence  seems 
to  stand  as  follows:  In  Detroit,  with  a  small  range  of  low  velocities 
(the  greatest  about  3  ft.  per  sec.)  the  30-in.  pipe  line  shows  an  increas- 


0.07 
0.06 

■  — 1-      ■[■     1 

1     1 

I                              1 

H 

180"CURVESilN  2.09-lNCH 

BRASS  PIPE   LINE 

^RITHMIC    PLOTTING  OF   DATA  FROM  TAB 

*ol, 

^ 

Log/ 

E   77, 

5>^ 

!l 

Straight  Pipe, in  Feet  of  Water, 

Transactions  AM  SOC.  C.  E.,  VOL.  XLVll, 

1902,PAGES  318-319. 
EXPERIMENTS  BY  SAPH  AND  SCHODER 

•  Curve  >.'o.l,  niuliub=^9.58  l'ij)e  Diiinieteis 

o      ••         ••    2        '■          7.03      •' 

+     ••         •■3        ••         5.77      •' 

O      •■         ••    -1         ••          4.7fi      •• 

a      "         "5        "           4.;i3      ■• 

®      •■         "6        "           .">.8}      '• 

/ 

1 

.*! 

V 

® 

O 

^■'i 

1 

r 

4// 

) 

A  /  / 

"3 

i'/ 

//T^ 

t) 

of  Head  ove 

#7^ 

) 

Excess  Loss 

1               i         i 

+ 

/ 

/ 

f 

/' 

i,  1 

7 

® 

< 

D 
/ 

/ 
/ 
/ 
u 

1 
/ 
/ 
/ 

/ 

/ 

/ 
/ 

/ 

/ 
1 

■n- 

® 
a 

0.5      0.()     0.7    0.3  O.y  1.0 

Velocity,  in  Feet  per  Second 
Fig.  12. 

iiig  loss  of  head  with  increasing  radius  of  curvature  for  curves  with 

radii  between  5  and  25  diameters.     (The  other  pipe  lines,  12  and  16-in., 

have  only  short-radius  curves,  and  the  results  are  not  consistent.)     The 

writer's  experiments  on  G-in.  i)ipe,  with  velocities  up  to  16  ft.  per  sec. 


Papers.]  CURVE  RESISTANCE  IN  WATER  PIPE  443 

show  the  opposite.  His  8-in.  pipe  experiments  indicate  no  measurable 
excess  loss  of  head  for  bends  composed  of  a  series  of  small  deflections 
a1  consecutive  joints  in  ordinary  cast-iron  pipe,  and  his  6-in.  pipe  also 
shows  no  appreciable  excess  loss  for  several  small  deflections.  The 
quantitative  results  for  the  long  Detroit  30-in.  90°  curves  are  far  in 
excess  of  the  180°  2-in.  brass  curves,  the  writer's  6-in.  90°  curves,  the 
3,  4  and  G-in.  screw  elbows,  and,  as  above  stated,  are  contradicted  by 
the  no-excess  results  of  the  8-in.  long,  easy  curves. 

If,  then,  for  any  reason,  an  engineer  wishes  to  use  a  long,  easy 
curve,  or  a  series  of  small  deflections,  in  the  joints  between  straight 
pipes,  the  writer's  experiments  indicate  just  what  most  hydraulic  en- 
gineers have  assumed,  namely,  that  there  is  practically  no  differenco 
between  the  loss  of  head  due  to  a  long,  easy  curve  and  that  due  to  an 
equal  length  of  straight  pipe. 

No  calculations  on  the  basis  of  the  loss  of  head  per  foot  length  of 
curved  portion  have  been  made,  because  all  the  excess  loss  of  head 
probably  does  not  occur  in  the  curve,  an  unknown  part  of  the  loss 
taking  place  in  the  down-stream  tangent  in  the  region  where  the  ab- 
normal flow  returns  to  normal.  Partly  for  this  reason,  also,  it  was 
not  deemed  wise  to  attempt  any  correction  on  account  of  the  smaller 
diameter  of  the  6-in.  cast-iron  curves,  or,  indeed,  on  account  of  the 
variation  in  the  diametei"s  of  any  of  the  curves  from  the  mean  diam- 
eter of  the  straight  pipe.  Besides,  we  have  no  precise  knowledge  con- 
cerning the  effects  of  slight  sudden  enlargements  or  contractions,  such 
as  are  involved  in  these  experiments. 

The  writer  has  found  this  an  interesting  study,  and  would  gladly 
have  extended  the  experiments  to  other  sizes  of  pipes,  but  the  cost,  in 
time  and  money,  is  rather  large,  lie  would  suggest  the  desirability 
of  similar  studies  on  other  small  sizes,  and  of  many  further  experi- 
ments on  curves  in  existing  large  pipe  lines. 

The  writer  desires  to  acknowledge  his  indebtedness  to  Professor 
W.  B.  Gregory,  who  worked  with  him  throughout  the  6-in.  pipe  ex- 
periments, and  in  subsequent  Pitot  tube  investigations  in  the  course 
of  which  it  was  shown  tliat  normal  flow  prevailed  at  all  the  piezometers 
in  the  6-in.  pipe. 

After  preparing  this  paper,  the  writer's  attention  was  called  to  a 
record  of  experiments  on  90°  bends  in  3  and  4-in.  pipes.*    These  ex- 

*  Paper  No.  3679,  "  Loss  of  Pressure  in  Water  Flowing:  through  Straight  and  Curved 
Pipes,"  by  Arthur  William  Brightmore.  M.  Inst.  C.  E.,  Mhiiitex  nf  Procefdinyx,  Inst.  C.  E., 
Vol.  CLXIX.  mw-ltHjT,  p.  393. 


444 


CURVE  RESISTANCE  IN  WATER  PIPE 


[Papers. 


periments  covered  a  right-angled  elbow  and  right-angled  bends  having 
radii  equal  to  2,  4,  G,  8,  10,  12,  and  14  diameters  in  ;3-in.  pipe,  and  the 
same,  excepting  the  last  two,  in  4-in.  pipe. 

The  arrangement  resembled  that  by  the  writer  except  that  the 
down-stream  piezometer  was  located  rather  close  to  the  curves,  being 
6  ft.  8  in.,  or  27  diameters,  distant  for  the  3-in.  pipe,  and  from  5  ft. 
to  6  ft.  1i  in.,  or  from  15  to  19  diameters,  distant  for  the  4-in.  pipe. 
The  length  of  straight  pipe  up  stream  from  the  up-stream  piezometer 
was  7  ft.  for  both  the  3  and  4-in.  pipes. 

Fig.  13  is  a  reproduction  of  Mr.  Brightmore's  plotting  of  the  re- 
sults of  his  experiments.     There  is  a  striking  similarity  between  the 


iD  OD  8D  IOC 

Radius  of  Beml,  in  Diameters. 
Fig.  13. 


shape  of  his  curves  and  those  by  the  writer  in  Fig.  3  (which  corre- 
sponds to  Fig.  13).  The  hump  in  the  curves  between  G  and  8  diam- 
eters appears  in  both  Figs.  3  and  13.  The  quantitative  results  are  not 
readily  compared.  For  Mr.  Brightmore's  4-in.  pipe  the  straight  pipe 
(rusted  cast  iron)  liad  a  coefficient  of  47.5  in  the  formula,  V  =  C  ^^  R  is. 
For  the  3-in.  pipe  (galvanized)  the  coefficient  was  65  to  70  for  V, 
ranging  from  3  to  11  ft.  per  sec.  In  the  writer's  experiments,  the  co- 
efficient for  the  G-in.  wrought-iron  pipe  was  119  to  125  for  V  ranging 
from  3  to  IG  ft.  per  sec.  Mr.  Brightmore  purposely  allowed  the  pipes 
and  curves  to  become  rusted,  but,  he  states,  not  tuberculated.  The 
foregoing  figures    indicate   that  he  was   working  witli   much    rougher 


Papers.]  CURVE  RESTSTANOE  IN  WATER  PTPE  445 

pipes  than  the  writer  used,  niid  tlie  quantitative  values  shown  on  Fig. 
13  indicate  the  same  in  comparison  with  Fig.  3. 

It  is  evident  that  further  experiments  are  desirable  before  precise 
laws  can  be  stated,  although  the  qualitative  results  by  Mr.  Bright- 
more  and  the  writer  agree  in  indicating  a  decreasing  loss  of  head  for 
an  increasing  radius  of  curvature. 


Vol.  XXXIV.  MAY,  1908.  No.  5. 


AMERICAN  SOCIETY  OF  CIVIL  ENGINEERS 

I  N  S  T  1  T  U  T  E  D    1  8  5  2 


PAPERS  AND  DISCUSSIONS 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced 
in  any  of  its  publications. 


NOTES  UPON  DOCKS  AND  HARBORS. 


By  Luther  Wagoner,  M.  Am.  Soc.  C.  E. 
To  BE  Presented  September  2d,  1908. 


During  the  year  1907  the  writer  visited  the  principal  ports  of 
Europe  and  the  United  States  for  the  purpose  of  procuring  data  for 
the  preparation  of  a  report  upon  the  future  needs  of  San  Francisco 
in  the  matter  of  port  improvements. 

For  the  foreign  work  he  had  exceptional  facilities  for  observing 
completed  works  and  receiving  information  concerning  new  or  pro- 
jected work.  As  a  result,  he  obtained  a  large  quantity  of  technical 
literature,  maps,  plans,  and  photographs,  as  well  as  notes,  and,  be- 
lieving that  some  of  the  data  obtained  may  be  of  general  interest  to 
the  profession,  he  lias  prepared  the  following  paper. 

Comparison  of  European  with  American  Harbors. 
A  striking  difference  in  the  ground  plans  of  port  works  is  at  once 
apparent  to  the  visiting  engineer.  The  development  by  piers  or 
jetties,  like  that  of  San  Francisco  or  New  York,  has  no  parallel.  Gen- 
erally speaking,  the  European  idea  is  one  of  enclosed  basins,  with  or 
without  locks,  as  tidal  conditions  may  require.  In  the  Mediterranean, 
where  the  tide  ranges  from  10  to  20  in.,  there  is  usually  no  protection 

Note. — These  papers  are  issued  before  the  date  set  for  presentation  and  discussion. 
Correspondence  is  invited  from  those  who  cannot  be  present  at  the  meeting,  and  may  be 
sent  by  mail  to  the  Secretary.  Discussion,  either  oral  or  written,  will  be  published 
in  a  subsequent  number  of  PiorerdiiKjs.  and.  when  finally  closed,  the  papers,  with 
discussion   Iti  full,  will  be  iiublished   in  Tra^isactioiis. 


Papers.]  NOTES  UPON  DOCKS  AND  HARBORS  447 

from  the  sea,  so  a  mole  or  breakwater  is  built  and  behind  it  a  safe 
harbor  is  created  by  dividing  up  the  protected  area  into  basins 
separated  by  solid  filled  piers,  usually  from  300  to  400  ft.  wide.  Upon 
these  piers  there  is  first  a  space  of  from  20  to  25  ft.  for  cranes  and  one 
railway  track.  Next  are  the  sheds,  from  75  to  130  ft.  wide;  then  there 
is  an  open  space  between  the  sheds  for  two  railway  tracks,  and  a  wagon 
road  between  them.  At  Naples  it  was  necessary  to  build  such  a  wall 
in  110  ft.  of  water;  at  Genoa,  in  depths  of  from  50  to  70  ft.,  and  at 
Marseilles  from  60  to  75  ft. 

At  cities  like  Antwerp,  Rotterdam,  Bremen,  and  Hamburg,  on 
tidal  rivers,  the  problem  has  been  one  of  obtaining  the  desired  area  by 
dredging  out  basins  and  enclosing  them  with  quay  walls,  leaving  suffi- 
cient space  for  railway  connections,  sheds,  warehouses,  and  roads,  the 
whole  being  arranged  so  as  to  facilitate  business.  In  this  matter  it  is 
specially  noteworthy  that  railway  connections  have  been  provided  at 
all  points,  and,  generally  speaking,  freight  can  be  transferred  directly 
from  the  ship  to  the  car  or  vice  versa,  thus  avoiding  delay  and  extra 
handling. 

Liverpool,  which  is  essentially  a  receiving  and  forwarding  port, 
has  perhaps  the  best  arrangement  of  railways  and  stations.  In  a 
length  of  6  miles  of  water  front  there  are  ten  or  more  great  railway 
freight  stations,  all  being  just  at  the  rear  of  the  docks.  In  point  of 
efficiency  of  belt-railway  service,  there  is  nothing  at  present  in  the 
United  States  that  is  comparable  with  that  of  any  first-class  European 
port.  Within  the  City  of  Philadelphia  there  are  more  railway  lines 
connecting  manufacturing  establishments  with  the  various  roads  than 
in  any  other  city  of  the  same  area,  but,  on  the  other  hand,  the  rail- 
ways own  and  control  about  80%  of  the  available  water  front.  New 
York  has  no  belt-railway,  and  it  is  doubtful  if  it  will  ever  have  one. 
as  the  price  of  land  required  for  it  is  prohibitory.  San  Francisco  is 
fortunate,  in  this  respect,  as  the  State  owns  the  water  front,  along 
which  thei"e  is  a  street  200  ft.  wide,  and  upon  which  there  is  a  State 
belt-railway.  If  the  recent  plans  for  the  port  improvements  are  carried 
out,  the  new  street  along  the  water  front  will  have  a  width  of  350  ft., 
upon  which  there  will  be  ample  room  for  belt-roads,  warehouses,  and 
railway  freight  stations,  as  well  as  street  cars  and  other  vehiciilar 
traffic. 

In  Europe  the  systematic  planning  of  new  work  is  especially  note- 


448  NOTES  UPON  DOCKS  AND  HARBORS  [Papers. 

worthy.  For  example,  in  Antwerp,  a  broad,  comprehensive  scheme  of 
port  enlargement  has  been  carefully  prepared,  and  is  of  such  magni- 
tude that  it  will  require  many  years  for  its  execution.  The  Board  of 
Control  has  acquired  the  lands,  and  planned  the  roads  and  rail  con- 
nections long  in  advance  of  actual  needs,  thus  permitting  its  orderly 
execution  without  regard  to  any  vested  interests.  As  the  city  grows 
up  around  the  new  port,  it  will  not  have  to  make  expensive  changes, 
Similarly,  Rotterdam  has  planned  a  gigantic  basin  where  it  is  in- 
tended to  dredge  650  acres  to  a  depth  of  40  ft.  and  deposit  the  soil  on 
the  low  lands  below.  It  is  estimated  that  the  completion  of  this  work 
may  require  from  twenty  to  thirty  years.  All  the  larger  ports  in 
Europe  are  planning  and  executing  systematic  extension  of  their 
facilities  to  hold  their  present  and  secure  a  share  of  expected  increased 
trade.  Coincident  with  such  work,  much  attention  is  given  to  making 
the  city  attractive,  a  place  where  one  would  like  to  live,  and  where  a 
visitor  would  like  to  go  again.  They  have,  in  general,  a  good  ad- 
ministration, and  are  able  to  select  and  keep  employed  men  of  ability 
to  administer  the  public  utilities.  As  a  rule,  they  look  further  aheavl 
than  Americans;  in  other  words,  they  think  more  before  taking  action. 

Statistical  Chart. 

Among  the  duties  imposed  upon  the  writer  was  the  reqiiest  to  try 
and  forecast  the  amount  of  the  future  commerce  of  the  port  of  San 
Francisco,  and  plan  improvements  ample  for  such  purpose,  say  fifty 
years  hence.  The  method  of  investigation  and  the  results  are  shown 
upon  the  diagram,  Plate  XL VIII.  After  a  preliminary  study  of  the 
data,  it  was  found  that,  owing  to  the  rapid  increase  in  the  quantities 
platted,  and  the  natural  irregularities  of  the  subject,  the  ordinary 
method  of  showing  the  time  relation  was  not  suitable  for  prediction 
purposes,  and  the  logarithmic  method  was  used.  In  the  diagram,  Plate 
XLVIII,  the  year  1700  is  zero,  1800  is  100,  and  1900  is  200,  and  the 
logarithms  of  these  numbers  were  used  for  the  time  scale.  Four  vor- 
tical scales  of  1  to  10  were  drawn,  so  as  to  cover  all  the  data  used 
without  confusion  of  lines.  In  such  a  diagram  an  inclined  line  de- 
notes an  exponent;  for  example,  the  average  line  drawn  through 
"Value  of  Merchandise  Exports  and  Imports,  TTnit(>d  States,"  is  the 
graphical  representation  of 

Average  value  =  0.0000244   (Year,  ITOO)*'"" 


♦ 

Foldout 

Here 
♦  ♦ 

♦ 


:  XLVIII. 

M.  SOC.  0.  E. 
,  1 908. 
NER  ON 
D  HARBORS. 

_1840      \~ 

=19p( 

) 

— b 

— t 

f. 

, 

-z 

1 

— ^. 

— « 

] 

•■r 

f. 

sir 

^ 

^ 

— 1- 

9 

^^ 

~ 



_ 

■ 

'      *  /• 

\ 

■";•  * 

■    * 

'     ■  '» 

) 

■ '  -r 

1 

-.-»■ 

■ 

I 

:■':'•. 

;i 

5- 

■■  ■ '  f 

-  * 

■:'; 
.'■':■ 

'■'■\ 
•■.;» 

'■■A 

,  > 

•.1 

•  .•  ■  ; 

'•■!'" 

2- 

] 

040 

1 

9p0 

I 

1 

Papers.]  NOTES  UPON  DOCKS  AND  HARBORS  449 

For  Hamburg's  tonnage,  the  exponent  is  about  10,  and  for  the  popu- 
lation of  the  United  States  the  exponent  is  about  4.  The  degree  of 
accuracy  of  this  method  of  forecasting  can  be  readily  seen  by  going 
back,  say,  to  1870  or  1880,  and  projecting  ahead  to  the  present  time. 
The  application  of  logarithmic  platting  to  such  purposes  is  believed  by 
the  writer  to  be  new. 

It  is  commonly  held  by  experts  that  the  production  of  pig  iron  and 
steel  is  a  good  financial  barometer.  The  production  of  iron  and  steel 
(not  shown  on  Plate  XLVIIl),  the  horse-power  used  in  the  United 
States  manufactures,  and  the  total  bank  deposits  in  the  United  States 
are  three  curves  which  can  be  almost  exactly  superimposed  by  moving 
them  vertically  into  position;  in  other  Vv'ords,  they  have  a  common  ex- 
ponent or  law  of  increase. 

Some  very  interesting  conclusions  may  be  drawn  from  this  diagram. 
For  example,  up  to  the  year  1900,  London  and  Liverpool  tonnage  were 
moving  at  a  common  rate  of  growth.  London  then  needed  port  im- 
provements, but  expended  its  energy  in  discussion  (and  a  very  thorough 
one  it  was),  while  Liverpool  deepened  and  extended  its  docks;  the  re- 
sult can  be  plainly  seen  upon  the  diagram  after  1900.  The  diagram 
also  shows  that  the  exponent  of  increase  of  population  in  the  United 
States  is  about  4;  exports  and  imports  of  the  United  States,  about  G; 
world's  commerce,  about  G;  United  States  bank  deposits,  about  10;  the 
horse-power,  and  manufacture  of  pig  iron,  steel,  etc.,  about  11  to  13; 
consequently,  per  capita,  this  means  a  rapid  increase  of  business  for 
the  engineer,  because  the  increase  in  such  activities,  referred  to  a  time 
relation,  is  measured,  per  capita,  by  the  difference  in  exponents. 

The  data  relating  to  steamships  are  quite  interesting.  They  were 
compiled  from  a  valuable  report,*  by  Elmer  L.  Corthell,  M.  Am.  Soc. 
C.  E.  Attention  is  called  to  the  very  rapid  and  uniform  increase  in 
the  average  tonnage  of  vessels,  about  841  tons  in  1873,  1  95,5  tons  in 
1903,  and  now  about  2  300  tons. 

The  draft  (loaded)  of  a  vessel  can  be  expressed  by  an  equation  in 
this  form : 

Draft  =  K   (length  X  breadth)^, 
in  which  K  ranges  from  1  for  small  vessels  to  0.88  for  those  of  the 
largest  type.     There   is  no  doubt  that  K  would  be  uniformly   taken 


*  To  the  Tenth  Congress,  Milan,  1905,  Permanent  International  Association  of  Naviga- 
tion Congresses. 


452  NOTES  UPON  DOCKS  AND  HARBORS  [Papers. 

water;  they  were  then  lowered  and  put  down  by  jetting  about  15  ft. 
into  the  sand,  care  being  taken  to  have  the  concrete  filling  at  all  times 
above  the  surface  of  the  water.  There  are  no  reinforcing  rods,  and, 
when  the  outer  galvanized  cylinder  fails,  the  structure  must  depend 
upon  the  tensile  strength  of  the  concrete  to  resist  the  lifting  action 
of  the  waves,  which,  owing  to  the  exposed  position  of  the  pier,  may  be 
quite  severe  during  a  storm. 

The  Wear  of  Concrete. 

Most  of  the  quay  walls  observed  in  Europe  are  faced  with  rubble  or 
ashlar.  In  Belgium,  Holland,  and  Germany  they  have  a  rubble  facing 
of  hexagonal  basalt  blocks  about  2  ft.  deep,  to  prevent  wear.  In  the 
Albert  Dock,  London,  a  concrete  non-faced  quay  wall  has  been  in  use 
about  30  years,  and,  having  been  svibjected  to  much  buffeting  from 
lighters  as  well  as  ships,  it  has  worn  away  about  2  in. 

Careful  inspection  was  made  as  to  a  possible  action  at  or  near  the 
water  line  due  to  freezing,  or  wave  action,  of  both,  but  nothing  note- 
worthy was  seen.  However,  at  Baltimore  there  is  a  noticeable  excep- 
tion, for,  on  certain  bridge  piers,  and  for  a  vertical  range  of  18  in., 
the  concrete  has  disintegrated  to  a  depth  of  several  inches  about  at 
the  ordinary  water  line.  Aside  from  the  affected  part  of  the  concrete, 
which  was  covered  with  a  vegetable  growth  more  dense  than  in  the 
lower  unaltered  part,  nothing  unusual  was  observed.  The  concrete 
above  and  below  the  affected  zone  is  good. 

Various  theories  have  been  advanced  to  account  for  the  decomposi- 
tion. The  most  plausible  one  is  that  with  a  small  tidal  range  there 
is  a  destructive  action  by  the  waves  lapping  the  affected  zone,  and  this, 
perhaps,  is  assisted  by  ice  action.  None  of  the  theories  suggested, 
when  weighed  and  considered  in  reference  to  similar  structures  else- 
where, appears  to  the  writer  as  tenable.  Believing  it  worthy  of  investi- 
gation, the  matter  has  been  reported  to  the  United  States  Geological 
Survey,  with  a  request  for  an  investigation  and  report. 

Concrete  Caissons. 
In  Europe  extensive  use  is  being  made  of  hollow  concrete  caissons, 
both  plain  and  reinforced,  for  breakwaters  and  quay  walls.  The 
structure  is  towed  into  position  and  sunk,  after  which  the  hollow  cells 
are  filled.  Some  of  those  used  for  biu^ikwaters  weigh  more  than  5  000 
metric  tons. 


Papers.]  NOTES  UrON  DOCKS  AND  HARBORS  453 

At  Rotterdam,  caissons  131.2  ft.  long,  and  having  a  width  of  32  ft. 
at  the  base  and  16  ft.  at  the  top,  and  43  ft.  high,  were  being  used.  A 
middle  division  wall  through  the  length,  and  nine  cross-walls,  divido 
the  caisson  into  twenty  cells.  Four  such  caissons  were  built  at  the 
same  time  in  an  improvised  dry  dock.  The  first  step  was  the  prepara- 
tion of  a  base,  about  2  ft.  thick  and  32  by  131.2  ft.,  well  reinforced, 
and  in  this  were  embedded  the  vertical  rods  for  the  walls.  The  ex- 
ternal side  walls,  about  14  in.  thick  over  the  base,  were  carried  up  witli 
a  batter.  When  the  caisson  walls  were  up  to  about  five-eighths  of  their 
final  height,  the  gates  were  opened  and  the  caissons  were  floated  out  to 
a  place  in  the  harbor  where  they  were  secured  to  mooring  piles.  There 
they  were  completed,  meanwhile  being  afloat  for  one  or  two  months. 

As  there  are  streaks  of  peat  in  the  soil  at  Rotterdam,  the  bad  parts 
are  removed  by  a  dredge,  and  then  the  dredged  cut  is  filled  with  sand 
at  least  6  ft.  deeper  than  the  base  of  the  caisson.  Then  the  caisson  is 
towed  into  place,  and,  by  means  of  a  tongue  and  groove  on  the  ends, 
the  floating  mass  is  brought  into  alignment,  the  free  end  being  con- 
trolled by  tackle.  Next,  by  opening  valves,  the  caissons  are  sunk  on 
the  prepared  bed  of  sand,  after  which  the  water  is  pumped  out  of  the 
front  row  of  cells  and  these  are  filled  with  concrete;  the  rear  row  of 
cells  is  filled  with  sand  to  save  expense. 

In  the  older  construction,  the  site  was  dredged,  then  a  brush- 
mattress  facing  was  placed,  and  this  was  allowed  to  stand  for  one  or 
two  years  to  secure  thorough  settlement  of  the  mass  on  and  to  the 
rear  of  the  mattresses.  Afterward  wood  piles  were  driven  through  the 
mattress  and,  by  using  a  special  diving  bell,  were  cut  off  and  capped 
below  low  water;  then  they  were  decked  with  wood  or  reinforced  con- 
crete upon  which  was  built  the  quay  wall.  The  floating,  reinforced 
concrete  caisson  method  was  stated  to  cost  less  per  linear  foot  than 
for  piles  decked  with  concrete  and  more  than  for  piles  decked  with 
wood.  Practically,  the  cost  may  be  said  to  be  the  same,  with  the  de- 
cided advantage  of  a  nearly  monolithic  wall. 

Where  rock  is  convenient,  this  method  might  be  used  with  ad- 
vantage :  Having  made  the  dredged  cut,  next  place  along  the  front  line 
of  the  caisson  one  or  more  rows  of  piles,  which  might  be  driven  to,  or 
cut  off,  say,  2  ft.  below,  the  grade  of  the  bottom  of  the  caisson;  and 
then  rock  fill  to  grade;  the  object  of  the  piles  covered  by  rock  would 


454  NOTES  UPON  DOCKS  AND  HARBORS  [Papers. 

be  to  prevent  any  rotation  of  the  concrete  block  around  the  outer  toe 
due  to  a  thrust  from  the  shore  side.* 

COMPRESSOL. 

In  Paris  the  writer  witnessed  a  demonstration  of  the  Compressol 
method  of  preparing  foundations.  By  a  sort  of  pile-driver,  a  heavy 
conical  weight  is  dropped  repeatedly  upon  the  soil,  and,  when  the  de- 
sired depth  is  obtained,  small  stones  are  dropped  into  the  hole,  and, 
by  special  forms  of  conical  weights,  are  forced  down  and  out  into  the 
soil,  after  which  concrete  is  rammed  into  place  by  the  same  means. 
A  completed  pile  will  generally  be  1  m.  in  diameter  and  have  a  bulb  at 
the  base. 

One  contractor  in  Belgium  has  eighteen  of  these  machines  at  work. 
There  are  many  places  where  such  a  system  might  be  used;  its  special 
value  would  appear  to  be  in  a  firm  soil  of  loam  requiring  piling,  and 
where  the  pile  heads  would  be  above  the  permanent  water  plane. 
AVhether,  in  point  of  economy,  it  presents  any  advantages  over  some 
of  the  patented  American  systems  is  not  known  to  the  writer.  It  is 
asserted,  by  those  advocating  the  Compressol  system,  that,  owing  to 
the  thorough  compression  given  to  the  grovmd,  both  laterally  and  ver- 
tically, combined  with  the  mushroom-shaped  base  of  the  pile,  it  is 
capable  of  sustaining  two  or  three  times  as  much  load  as  piles  used  in 
American  systems.  Certainly  there  must  be  considerable  merit  in  it, 
otherwise  it  would  not  be  used  so  extensively. 

Wood  Piles  Driven  at  an  Angle. 

At  Bremen,  and  notably  at  Bremerhaven,  it  is  the  practice  to  con- 
struct much  of  the  new  work  in  the  dry.  The  area  to  be  enclosed  is 
stripped,  by  land  dredges  and  cars,  to  3  or  4  ft.  below  low  water,  and 
is  kept  dry  by  pumps.  Along  the  proposed  line  of  quay  wall,  two 
single-rail  tracks  are  laid,  about  33  ft.  apart;  these  serve  to  carry  a 
pile-driver  which  traverses  a  carriage  supported  by  the  two  rails.  Thft 
pile-driver  is  arranged  to  swivel  in  two  directions;  thus  the  driver  can 
be  placed  with  great  precision  and  dispatch,  and  piles  can  be  driven 
at  any  desired  batter.  It  is  the  practice  to  select  long  piles,  and  first 
drive  each  tenth  bent  of  piles.     The  ways  are  marked  with  a  metric 

*  A  good  illustrated  technical  description  of  the  work  at  Rotterdam  may  be  found  in 
De  Ingenicur,  July  20th.  1907;  The  Hague,  Holland.  This  has  not  yet  been  translated  into 
English. 


Pi'I'f'^']  NOTES  UPON  DOCKS  AND  IIAKBORS  455 

scale,  and  an  attendant  records  the  position  of  the  pile  at  each  fifth 
or  tenth  blow.  From  these  data  piles  of  suitable  length  are  selected 
for  the  intervening  nine  bents,  and  a  similar  record  of  driving  is 
kept.  Should  one  or  more  piles  in  a  group  settle  too  much  during 
the  last  ten  blows,  a  longer  pile  is  driven  in  the  bent  or  in  the  ad- 
joining bents  to  give  additional  bearing  power.  The  bents,  when 
driven,  batter  about  1  on  5,  like  the  letter  A.  At  the  cross  of  the  A, 
two  strong  timbers  are  bolted  to  the  framed  piles,  and  longitudinal 
wales  are  bolted  to  the  i)iles  and  side  pieces,  and  a  6-in.  wood  floor 
completes  the  foundation.  Upon  this  fovmdation  a  quay  wall  of  rubble- 
faced  concrete  is  built,  and  is  bonded  to  the  rear  piles  by  tension  rods, 
after  which  the  area  is  opened  and  excavated  to  the  full  depth  by 
dredges.  A  construction  of  this  kind  resists  most  effectively  the  thrust 
from  the  landward  mass  of  earth,  the  outer  piles  being  compressed  and 
the  rear  piles  in  tension. 

Mr.  Claussen,  Dock  Engineer  at  Bremerhaven,  says  that  he  con- 
siders 15  tons  per  pile  a  safe  load  when  used  in  tension.  A  number 
of  long  walls,  such  as  the  sides  of  a  dry  dock,  quay  walls,  and  locks 
in  use  for  ten  years,  were  remarkably  straight,  and  offered  strong  evi- 
dence as  to  the  value  of  this  system  of  construction.  This  could  not 
be  used  if  limnoria  or  teredo  were  present,  unless  the  piles  in  the  outer 
row  were  covered  by  a  concrete  wall;  but  it  might  have  application 
for  mooring  bits,  etc.,  where  the  piling  is  protected,  and  perhaps  would 
be  advantageous  where  it  is  required  to  erect  a  temporary  bulkhead 
and  load  the  ground  landward  for  a  year  or  more,  so  as  to  consolidate 
it  before  commencing  the  permanent  construction  of  a  quay  wall. 

Cranes. 

In  nothing  is  the  difference  between  the  United  States  and  Europe 
so  marked  as  in  the  non-use  here  and  the  general  use  there  of  power 
cranes,  usually  hydraulic,  but  often  electrically  operated. 

It  is  quite  common  for  a  merchant  to  visit  Europe  and,  having 
noted  the  many  excellent  things  to  be  seen,  quite  naturally  think  that 
among  the  improvements  required  in  the  United  States  are  cranes. 

In  Europe  several  thousand  cranes  are  installed;  upon  an  average 
there  is  one  crane  to  283  ft.  of  quay  wall,  and  their  usage  is  quite 
variable. 


456  NOTES  UPON  DOCKS  AND  IIAKBORS  [Papers. 

At  Marseilles,  in  1903,  thirty-four  hydraulic  cranes,  having  a 
capacity  of  2  750  lb.,  worked  121  days  per  year  for  each  crane,  on  a 
J)-hr.  basis,  and  averaged  25.2  loads  per  hour.  Ten  cranes  of  double 
power,  but  working  at  one  metric  ton,  worked  115  days;  and  working 
at  3  tons,  16  days  each  per  year. 

The  whole  number  of  loads  was  1  116  980,  which  gave  an  income  of 
166  312  francs,  or  2.92  cents  per  load.  The  average  income  of  a  crane 
was  about  $745  per  year,  which  includes  the  power  and  the  crane 
operator.  This  port  has  perhaps  a  more  intensive  use  of  cranes  than 
those  farther  north,  where  generally  only  one  crane  out  of  four  or  five 
is  observed  to  be  working. 

At  London  it  is  alleged  that  cranes  do  not  pay  interest  upon  their 
cost,  but  the  ship  owners  insist  that  the  dock  owners  have  them  and 
do  not  use  them,  except  perhaps  for  a  small  part  of  a  cargo. 

It  is  possible  that  the  crane  idea  is  a  survival  from  the  days  of 
sailing  ships,  when  they  were  first  introduced  and  were  really  re- 
quired; next  they  were  copied  by  other  places,  and  by  sheer  inertia 
dock  owners  persist  in  having  them.  The  people  of  the  United  States 
are  quick  to  seize  and  appropriate  a  good  idea,  and  the  fact  of  the 
non-use  of  cranes,  compels  a  strong  belief  that  the  appliances  used — 
the  ships'  tackle  and  the  stevedores'  hoists — are  ample.  It  has  been 
asserted  that  a  difference  in  the  nature  of  the  business  done  in  the 
different  countries  is  responsible  for  the  general  use  of  cranes  abroad; 
but  this  does  not  appear  to  be  a  reasonable  view  of  the  matter. 

The  writer  does  not  wish  to  give  the  impression  that  cranes  are  not 
useful;  on  the  contrary,  he  believes  that  a  partial  adoption  of  the  plan 
in  the  United  States — to  the  extent,  at  least,  of  having  wharf  cranes 
which  would  serve  to  lift  anything  in  excess  of  the  capacity  of  a 
ship's  tackle — might  prove  useful. 

At  Liverpool  many  cranes  are  supported  upon  the  top  of  the  ware- 
house front  wall  and  a  rail  upon  the  peak  of  the  roof.  This  is  a  very 
excellent  disposition,  because  the  crane  is  always  out  of  the  way. 

Dock   Strikes. 

There  is  and  has  been  much  troiiblo  at  many  ports  in  Europe 
from  labor  strikes,  and,  upon  the  whole,  these  arc  probably  worse  than 
any  that  have  happened  in  America. 

It  is  difficult  for  a  stranger  to  form  an  accurate  estimate  of  this 


Papers.]  K0TK8   Ul'ON  DOCKS  AND  llARBOUS  457 

subject,  because  he  hears  various  versions  of  the  cause  and  nature  of 
the  trouble. 

The  writer  questioned  the  officials  of  the  ports  visited,  and,  as  n 
check,  obtained  the  views  of  the  marine  underwriters  and  sometimes 
the  Jesuit  Fathers,  who  are  usually  in  a  position  to  estimate  the 
troubles  impartially.  The  general  unrest  appears  to  have  below  it  a 
raising  of  the  standards  of  living,  for  the  cost  of  living  and  the  wants 
of  the  laborer  have  increased  faster  than  his  wages.  On  the  other 
side,  the  employers  assert  that,  in  view  of  the  serious  competition  be- 
tween ports,  a  small  amount  of  extra  cost  will  cause  a  diversion  of 
business,  therefore  they  oppose  an  increase  of  wages. 

The  nature  of  the  work  is  irregular,  there  are  periods  of  great 
activity  followed  by  lessened  opportunity  for  work.  Genoa  has  been 
greatly  troubled  in  the  past  by  strikes,  and  the  present  port  governing 
board,  which  is  closely  modelled  on  the  lines  of  the  Liverpool  Dock 
Trust,  has  assumed  that  it  has  the  power  to  settle  such  questions  by 
creating  a  permanent  force  of  laborers  who  perform  any  sort  of  service, 
from  discharging  cargo  to  road  making,  if  required.  The  board  takes 
on  extra  men  for  short  periods  to  cover  emergency  cases.  The  laborers 
receive  less  pay  than  men  engaged  by  private  employers,  and  it  is  said 
that,  on  account  of  a  strong  union  organization,  they  do  less  work. 
The  ovitside  criticism  was,  that  the  laborers  practically  dictated  hour'3 
and  terms  to  the  dock  board.  The  experiment  is  an  interesting  one, 
and  its  outcome  will  be  a  matter  of  interest. 

During  the  writer's  visit,  a  mild  strike  was  in  progress  at  Le  Havre, 
and  at  Antwerp  a  severe  one  which  required  the  importation  of  some 
2  000  English  strike  breakers,  who,'  for  their  protection  from  assault, 
were  housed  on  vessels  in  the  harbor.  In  August  the  strike  culminated 
in  burning  the  timber  yards  and  required  calling  out  the  troops  in 
order  to  save  the  city  from  fire  and  quell  the  rioters. 

To  the  writer,  the  wages  seemed  to  be  too  low,  but  unless  concerted 
action  were  taken  by  all  the  competing  ports,  it  would  be  difficult  to 
effect  a  raise.  It  does  not  appear  that  the  form  of  government  has 
any  very  decided  influence  upon  strikes  or  violence  arising  therefrom. 
The  striker  is  usually  a  voter,  and  some  one  in  authority  may  need 
his  vote.  About  the  only  safe  deduction  that  can  be  made  is  that  a 
small  aggressive  minority  comes  pretty  near  getting  all  it  desires. 


458  NOTES  UPON  DOCKS  AND  HARBORS  [Papers. 

Average  Data. 

The  following  average  data  concerning  foreign  ports,  compiled 
from  a  paper,*  entitled  "Results  of  Investigation  Into  Cost  of  Ports 
and  Their  Operation,"  by  Mr.  Elmer  L.  Corthell,  and  corrected  by  the 
writer  in  a  few  particulars,  presents  in  a  summarized  form  informa- 
tion of  considerable  interest: 

The  ports  included  are  London,  Liverpool  and  Birkenhead,  Glas- 
gow, Bristol,  Hamburg,  Rotterdam,  Le  Havre,  Dunkirk,  Bilbao,  Ant- 
werp, Bremen  and  Bremerhaven,  the  Tyne  Ports,  Marseilles,  Amster- 
dam, Lisbon,  Bombay,  and  Buenos  Ayres. 

Total  cost  of  port  improvements  to  1906 $764  388  000 

Registered    tonnage,    entered    and    cleared    in    one    year 

(about  1905-1906)    185  652  000 

Goods   dealt   with   in   one  year,   1905-1906,   in   long  tons 

of  2  240  lb.,  approximately 143  000  000 

Gross  revenue  in  one  year  (about  1905-1906) 58  206  000 

Expenses             "      "       "           "          "         "     29  003  000 

Net  revenue       "      "       "           "         "        "     29  203  000 

Gross  revenue  per  registered  ton         "         "     31.4  cents 

Gross  revenue  per  long  ton                  "         "     40.6  cents 

Quayage  length,  in  feet 1 192  000 

Quayage  length,  in  miles 227 

Approximate  length  of  rail,  in  miles 930 

Length  of  rail  divided  by  length  of  quayage 4.11 

Ratio  of  area  of  sheds  to  quayage  (44  467  770  sq.  ft.  of 

sheds  to  1  067  120  ft.  of  quay) 41.1  to  1 

Average  weight  of  goods  dealt  with  per  year  per  linear 

foot  of  quay,  in  long  tons 120 

Percentage  of  gross  income  on  capital  cost 7.615% 

One  crane  to  each  283  ft.  of  quay  wall. 

General  Reflections  on  Commerce. 

The  greatest  factors  to-day  in  the  material  and  moral  development 
of  the  world  are  transportation  and  commerce.  By  their  agency,  people 
and  their  products  are  moved  from  a  region  of  a  lesser  to  one  of  a 
greater  use  and  demand.     They  are  the  greatest  of  all  the  civilizing 

*  Proceedings  of  Permanent  International  Navigation  Congresses,  Brussels,  1907. 


Papers.]  NOTES  UPON  DOCKS  AND  HARBORS  459 

agencies,  because  they  promote  an  exchange  of  thought  as  well  as  of 
commodities. 

The  growth  of  modern  commerce  is  closely  interwoven  with  the 
development  of  the  steam  engine,  railways,  and  electricity.  It  is  a 
question  of  power  and  its  applications,  and  its  present  enormous  di- 
mension is  largely  the  work  of  the  engineer.  Its  growth  has  been 
phenomenal,  and  is  ever  increasing;  it  is  far  more  rapid  in  its  rate  of 
increase  than  that  of  population,  which  means  increased  wants  upon 
the  part  of  the  people,  and  increased  ability  to  buy  and  to  enjoy.  Its 
future  is  a  question  of  great  philosophical  interest;  but,  until  the  people 
of  the  world  are  raised  to  the  general  level  of  intelligence  of  the  more 
favored  nations,  it  is  reasonable  to  believe  that  its  march  will  continue, 
and  that  day  is  so  far  distant  that  it  does  not  immediately  concern  the 
present  age.  Until  then,  the  signs  point  to  an  ever-increasing  scope  in 
the  functions  of  the  engineer. 


Vol.  XXXrV.  MAY,  1908.  No.  5. 


AMERICAN  SOCIETY  OF  CIVIL  ENCmEERS 

rNSTITUTED    185d 


PAPERS  AND  DISCUSSIONS 

This  Society  is  not  responsible,  as  a  body,  for  the  facts  and  opinions  advanced  in 
any  of  its  publications. 


THE  FLOOD  OF  INIARCH,  1907, 

IN  THE  SAORAIMENTO  AND  SAN  JOAQUIN  RIVER 

BASINS,  CALIFORNIA. 

Discussion.* 


By    Messrs.   Lcttirr   Waoonrr,   IT.   IT.    Wadswortii,    and   George   L. 

DiLLMAN. 


Mr.  Wagoner.  LuTiiER  Wagoner,  M.  Am.  Soc.  C.  E.  (by  letter). — This  paper  is  an 
extremely  valuable  eontribution,  and  presents  the  main  facts  of  the 
flood  with  great  clearness.  The  opening  statement,  that  it  was  one  of 
the  most  destructive  floods  that  has  ever  occurred  in  California,  while 
probably  correct  in  a  financial  sense  (and  due  to  the  fact  that  there 
was  more  property  to  be  damaged  than  at  previous  floods),  implies  that 
it  was  about  the  greatest  flood  on  record.     The  authors  say: 

"It  is  doubtful  if  any  combination  of  causes  or  conditions  will 
ever  produce  a  larger  rate  of  delivery  of  water  to  this  valley  for  a 
4-diiy  ]>oriod  than  occurred  during  the  flood  of  March,  1907." 

The  writer  believes  that  it  would  be  unsafe  to  accept  this  statement 
as  a  basis  for  planning  reclamation  and  flood  prevention,  unless  it  is 
qualified  by  a  large  factor  of  safety.  It  is  generally  believed  that  the 
flood  of  1862  was  greater  in  volume  of  water  discharged  into  the  basins 
and  bay.  In  1890  the  writer,  while  engaged  upon  plans  for  the  La 
Grange  Dam  on  the  Tuolumne  River,  found  a  well-preserved  record  of 
the  18G2  flood  near  the  present  dam  and  70  ft.  above  the  bed  of  the 
stream.     The  record  was  in  the  shape  of  rounded  pieces  of  wood  and 

*  Continued  from  April,  1008.  Procceilinris. 


Papers.]  DISCUSSION  ON  CALIFORNIA  FLOOD  461 

bark,  fir,  pine,  tamarack,  and  juniper,  showing  that  these  pieces  came  Mr.  Wagoner, 
from  tlie  higher  regions.  They  were  found  in  a  talus  of  loose 
rock,  and  were  doubtless  carried  into  the  void  spaces  by  eddies  and 
lodged  there.  Almost  opposite,  and  across  the  river,  a  similar  deposit 
was  found,  and  at  almost  the  same  level.  This  led  to  a  search  along 
the  river  gorge  above,  where  several  similar  records  were  found.  Levels 
were  taken,  and  connected  with  several  cross-sections,  and  these,  com- 
bined with  the  known  high  water  at  La  Grange,  about  1  mile  below  the 
dam,  where  the  channel  is  wider  and  more  regular,  led  to  the  conclusion 
that  the  maximum  discharge  was  130  000  cu.  ft.  per  sec.  This  was 
based  on  the  slope  and  Ivutter's  formula  (n  =  0.040),  and  the  dam  was 
planned  to  be  able  to  discharge  that  volume  over  it;  this  corresponds 
to  a  run-off  of  8G.7  cu.  ft.  per  sec.  per  sq.  mile. 

In  1895  the  writer  found  a  similar  record  on  the  middle  fork  of  the 
American  River  near  Volcanoville,  from  which,  by  the  same  methods, 
a  run-off  greater  than  100  cu.  ft.  per  sec.  per  sq.  mile  was  deduced. 
(The  original  notes  of  both  the  foregoing  records  were  destroyed  in  the 
San  Francisco  fire,  two  years  ago.)  Estimates  have  appeared  in  print 
in  which  the  flood  flow  was  given  for  the  whole  basin  of  the  American 
River  at  250  000  cu.  ft,  per  sec,  and  even  more.  The  record  given 
above,  applied  to  the  whole  water-shed  above  Fair  Oaks,  would  give  a 
discharge  of  about  double  the  authors'  93  000  cu.  ft.  per  sec. 

While  it  may  be  true  that  a  flow  of  782  000  cu.  ft.  per  sec.  for  4 
days  may  not  be  exceeded,  there  are  two  points  to  be  considered.  The 
flow  from  the  San  Joaquin  region  might  occur  as  in  1862,  and  in 
combination  with  a  1907  flood  on  the  Sacramento,  in  which  case  the 
quantity  would  be  greatly  exceeded.  Again,  suppose  the  rivers  were 
leveed  in  accordance  with  the  plans  of  the  Engineering  Commission 
of  1904,  it  would  not  require  a  4  days'  sustained  flood  to  overtop  the 
levees,  and  the  probabilities  are  always  in  favor  of  the  shorter  but 
perhaps  more  intense  run-off. 

There  is  an  average  difference  of  a  month  in  the  melting  of  the 
snow  upon  the  Columbia  and  Snake  water-sheds,  yet  in  1900  it  melted 
on  each  at  the  same  time,  with  the  result  of  backing  up  the  Willamette 
and  flooding  Portland  to  a  depth  of  several  feet.  This  flood  was  sus- 
tained for  more  than  two  weeks. 

The  writer  concurs  in  the  conclusions  of  the  authors,  that  relief 
from  damage  by  floods  must  be  sought  in  storage  to  relieve  the  peak 
of  the  discharge.  Storage  of  debris  is  equally  important,  if  the  bed  of 
the  stream  and  navigation  interests  are  to  be  preserved,  and  future 
studies  should  be  upon  the  lines  of  effecting  both  water  and  debris 
storage. 

It  is  the  writer's  belief  that  this  can  best  be  accomplished  by  loose- 
rock  dams,  backed  with  earth  and  waste  on  the  up-stream  face  and 
with  the  down-stream  face  secured  to  the  mass  by  suitable  anchors,  so 


462  DISCUSSION  ON  CALIFORNIA  FLOOD  [Papers, 

Mr.  Wagoner,  fis  to  allow  the  passage  of  floods  over  the  crest  of  the  unfinished  dam 
during  construction.  When  completed  there  would  be  an  ample  spill- 
way around  the  dam,  so  that  it  could  never  be  overtopped.  Such  dams 
would  have  to  be  of  great  height,  from  400  to  500  ft.,  or  even  greater, 
because  they  would  usually  be  located  in  gorges,  and  it  might  require 
from  300  to  400  ft.  of  permanent  elevation  to  create  a  sufiicient  reser- 
voir area,  after  which  the  increase  of  storage  would  be  rapid.  The  only 
serious  objection  to  such  a  type  is  the  cost,  but  it  can  be  shown  that  in 
the  end  it  would  be  economical,  because  the  desired  regulation  could 
thus  be  obtained  (and,  incidentally,  the  storage  of  debris),  as  well  as 
power  and  irrigation. 

It  is  beginning  to  be  recognized  that  the  proper  treatment  of  this 
subject  is  a  serious  matter,  and  that  the  cost  may  reach  $100  000  000 
or  more.  From  the  analysis  presented  in  the  paper,  it  appears  that 
there  should  be  a  storage  of  about  3  000  000  acre-ft.  in  order  to  give  the 
required  relief  on  the  Sacramento  water-shed  alone.  Such  storage  could 
be  valued  as  follows:  (a)  Relief  of  peak  load  and  flood  prevention; 
(h)  storage  of  debris;  (c)  preservation  of  the  channel  of  the  river;  (d) 
irrigation;  and  (e)  power.  Wlien  all  these  possible  uses  are  admitted 
and  properly  valued,  it  can  readily  be  seen  that  a  high  cost  per  acre- 
foot  of  storage  is  permissible. 
Mr.  Wads-  H.  H.  Wadswortii,  M.  Am,  Soc.  C.  E.  (by  letter). — Although  dis- 
astrous to  the  agricultural  interests  of  so  large  an  area,  and  to  all  the 
transportation  lines  of  the  valley,  in  some  respects  it  may  be  said  that 
this  flood  occurred